document resume - eric · education. five speeches and four panel discussions dealt with computers....

ED 058 376

TITLE

\INSTITUTIONSPONS AGENCY

PUB DATENOTE

EDRS PRICEDESCRIPTORS

DOCUMENT RESUME

VT 010 830

Conference on the Use of Computers in MedicalEducation (Mlahoma City, Okla., Apr. 3, 4, and 5,

1968).Oklahoma Univ., Oklahoma City. Medical Center.Public Health Service (DREW), Washington, D.C. Bureau

of Health Manpower.Apr 68104p.

MF-$0.65 HC-$6.58*Computer Assisted Instruction; Computer BasedLaboratories; *Computer Oriented Programs;*Computers; Conference Reports; Group Discussion;*Medical Education; Medical Libraries; Speeches

ABSTRACTThe purpose of this conference was to explore what is

currently being done with computers in medical education, what future

goals should be established for computers, and how these goals should

be implemented in medical education. Five speeches and four panel

discussions dealt with computers in undergraduate medical education,

clinical medical education, continuing medical education, and medical

libraries. Also included are four broad-scope presentations, a

statement from the Surgeon General, and a progress report on computer

assisted instruction at the University of Oklahoma Medical Center.

(RC)

\/-leprinted with permission by the

U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE

National Institutes of Health

R.,

4

APRIL 3, 4, & 5, 1968 OKLAHOMA CITY, OKLA.

;4

PROCEEDINGS

Conference

U.S. DEPARTMENT OF HEALTH. EDUCATION& WELFARE

OFFICE OF EDUCATIONTHIS DOCUMENT HAS BEEN REPRODUCED

EXACTLY AS RECEIVED FROM THE PERSON ORORGANIZATION ORIGINATING IT. POINTS OF

VIEW OR OPINIONS STATED DO NOT NECES.

SARILY REPRESENT OFFICIAL OFFICE OF EOU.

CATION POSITION OR POLICY

on the Use of Computers in

Medical Education

Co-Sponsored by

U. S. Department of Health, Education and Welfare

Public Health Service

Bureau of Health Manpower

and

The School of Medicine

University of Oklahoma Medical Center

April 3, 4, and 5, 1968

University of Oklahoma Medical Center

Oklahoma City, Oklahoma

TABLE OF CONTENTSPage

Foreword5

James L. Dennis, M.D.

Program7

Participants9

Medical Technology and Social Change 11

J. Herbert Hollomon, Ph.D.

Conference Objectives14

Joseph M. White, M.D.

Statement of the Surgeon General 16

Excellence Buy or Beget? 17

Frank W. McKee, M.D.

Evaluation of the Use of Computers in Medical Education 22

Stephen Abrahamson, Ph. D.

EDUCOM A Progress Report 28

James G. Miller, M.D., Ph.D.

Presentation of Conference Subjects

Computers in Undergraduate Medical Education 37

Hilliard Jason, M.D., Ed. D.

Summary of Panel Discussion45

Inquiry, Computers and Medical Education 48

Lee S. Shulman, Ph.D.

Computers in Clinical Medical Education 53

Donald A. B. Lindberg, M.D.

Summary of Panel Discussion 57

Computers in Continuing Medical Education 64

George E. Miller, M.D.

Summary of Panel Discussion70

Computers in Medical Libraries 74

Martin M. Cummings, M.D.

Summary of Panel Discussion 93

Summary of Demonstration of C. A. I. System 100

William G. Harless

Closing Comments105

General Resolutions and Recommendations 105

FOREWORD

JAMES L. DENNIS, M.D.

University of Oldahoma Medical Center

Distinguished guests and colleagues, welcome to Oklahoma, to OklahomaCity and to the University of Oklahoma Medical Center campus.

Someone has described planning as the projection of a wedding betweenthe ideal and the practical. This is not easily accomplished in our complex society,but one of the catalytic agents that can make such a wedding possible is computerscience. Current pressures upon medical education are twofold. From one side,there is the explosion of knowledge and from the other side, the pressures fromsociety to fill the gaps in the health care system and to maintain the ongo4igquality of the physicians we graduate to prevent obsolescence. Computer scienceoffers perhaps the most attractive possibility for bringing order out of chaos.

When faced with the problem of relating medical education to societal needs,

one must be very concerned about the protection of academic excellence and

its future enhancement. The pressures to become involved in the delivery of

mass health services to relieve the shortages and deficits that exist in the nation

are tremendous. Many people now turn to the medical centers to solve their

problems and we cannot ignore our responsibilities. The situation is similar tothat of the chicken and the pig who were worried about their kind master who

was having a very hard time. The?, wondered what they could do to help. Thechicken said, "We can feed him,' and the pig asked, "Feed him what"? Thechicken responded, "We can feed him ham and eggs" to which the pig reactedimmediately, "For you that's only a commitment, but for me that's total involve-ment." Certainly, anytime a medical educational institution takes on the re-sponsibility of an undefined mass health service program, this means total in-volvement and with the same results that the pig anticipated. Therefore, wemust have a protected core of academic excellence, as is found in the researchand educational programs. Here is the fountainhead for new knowledge and newdevelopments. The development of a health center concept whose activities getprogramming closer to the public, provides communication from the research

bench and the academic desk by way of programs that permit an immediate

transmission of new knowledge and new technology to the public.

We thought that the health service problems on the one hand and the health

science challenges on the other were more than a single medical institution in astate such as Oldahoma could meet, without forming an alliance with others in

our community who have responsibilities for the health sciences. We went to out-standing citizens of this state and discussed the possibility of developing what

we now call the Health Sciences Foundation. This foundation, was formed not

*Vice Presidant for Medical Center Affairs and Dean, School of Medicine

5

to run the proposed Oklahoma Health Center but to aid and assist institutionswho were willing to build in geographic proximity to the University as a coreand to commit themselves to assisting the University in its mission of health man-power production and the pursuit of knowledge. The planning is well alongand buildings are under construction. The Health Center will have interfaces withinstitutions at city, county, state, and Federal health agency levels and withboth private and public agencies and institutions. Many central services andfacilities will be shared by all of these institutions. This means that funds forcapital investments that we receive at the Medical Center can go into labora-tories, classrooms and beds rather than into steam, refrigeration and laundries.The Oklahoma Health Center is coordinated by an operations committee com-prised of the heads of each of the institutions that will locate here. This commit-tee determines feasibility and recommends policy and procedures to our in-dividual governing boards. These are presentecl to the Foundation for implemen-tation. All of these activities are tied together by a rather ambitious computerprogram and, of course, the concept of reaching out into the state is very muchdependent upon computer technology and computer-based communication.

This conference, to explore the role of computers in medical education, is avery timely one because of the current examination into curricula, programs andfacilities for health ir anpower. Much should be learned from this opportunity forknowledgeable people to interact, and the discussion could have a significantimpact on the future of medical education.

8

r

PROGRAMWEDNESDAY, APRIL 3, 1968

Chairman Edward N. Brandt, Jr., M.D., Ph.D., Conference Director

WELCOME James L. Dennis, M.D., Vice President for Medical Center Af-fairs, University of Oklahoma Medical Center

MEDICAL TECHNOLOGY AND SOCIAL CHANGEJ. Herbert Hollomon, Ph.D., President-DesignateUniversity of Oklahoma

CONFERENCE OBJECTIVESJoseph M. White, M.D., Dean of Medical FacultyUniversity of Oklahoma School of Medicine

EXCELLENCE BUY OR BEGET?Frank W. McKee, M.D., Director of the Division of PhysicianManpower, U. S. Public Health Service

EVALUATION OF THE USE OF COMPUTERS IN MEDICAL EDUCATIONStephen Abrahamson, Ph.D., Director, Division of Research inMedical Education, University of Southern California, Schoolof Medicine

Panel DiscUssion by Faculty Members of the University of Oklahoma Medical

Center Involved in Computer-Aided Instruction and Demonstration of the CAI

System

Chairman: William C. HarlessPanel: Thomas N. Lynn, M.D., Preventive Medicine

and Public.HealthArthur W. Nunnery, M.D., PediatricsEugene D. Jacobson, M.D., PhysiologyRobert C. Duncan, Ph.D., Biostatistics andEpidemiologyJames A. Cutter, M.D., Anesthesiology

THURSDAY, APRIL 4, 1968Chairman: Joseph M. White, M.D.

EDUCOM A Progress ReportJames G. Miller, M.D., Ph.D., Principal Scientist,EDUCOM, and Academic Vice-President, ClevelandState University

PRESENTATION OF CONFERENCE PANEL SUBJECTSComputers in Undergraduate Medical Education

Hilliard Jason, M.D., Ed.D.Computers in Clinical Medical Education

Donald A. B. Lindberg, M.D.Computers in Continuing Medical Education

George E. Miller, M.D.Computers in Medical Libraries


7

f

CONCURRENT PANEL DISCUSSIONS

Computers in Undergraduate Medical Education

Chairman: Hilliard Jason, M.D., Ed. D.

Panel: Stephen Abrahamson, Ph.D.Alan Kaplan, M.D.Theodore Peterson, Ph.D.Lee S. Shulman, Ph.D.

Computers in Clinical Medical Education

Chairman: Dona 7d A. B. Lindberg, M.D.

Panel: Robert Greene% M.D.Mitchell SchorowJohn A. Starkweather, Ph.D.James Sweeney, Ph.D.

Computers in Continuing Medical Education

Chairman: George E. Miller, M.D.

Panel: William T. Blessum, M.D.Edward N. Brandt, Jr., M.D., Ph.D.Gerald Escovitz, M.D.Alexander Schmidt, M.D.

Computers in Medical Libraries

Chairman: Martin M. Cummings, M.D.

Panel: Robert DivettLeonard M. EddyIrwin Pizer

FRIDAY, APRIL 5, 1968

Chairman: Norman E. Tucker

SUMMARY AND REPORTS ON PANEL DISCUSSIONSComputers in Undergraduate Education

Hihiard Jason, M.D., Ed.D.Computers in Clinical Medical Education

Donald A. B. Lindberg, M.D.Computers im Continuing Medical Education

George E. Miller, M.D.Computers in Medical Libraries


RESOLUTIONS AND RECOMMENDATIONS

8

PARTICIPANTSSTEPHEN ABRAHAMSON, Ph.D., Director, Division of Research in Medical Education,

School of Medicine, University of Southern California, Los Angeles, California

WILLIAM T. BLESSUM, M.D., Medical Computer Facility, California College of Medicine,University of California at Irvine, Los Angeles, California

EDWARD N. BRANDT, JR., M.D., Ph.D., Director, Computer Facility, University of Okla-homa Medical Center, Oklahoma City, Oklahoma

MARTIN M. CUMMINGS, M.D., Director, National Library of Medicine, U. S. PublicHealth Service, Department of Health, Education and Welfare, Bethesda, Maryland

JAMES A. CUTTER, M.D., HecI, Department of Anesthesiology, University of OklahomaMedical Center, Oklahoma City, Oklahoma

JAMES L. DENNIS, M.D., Vice-President for Medical Center Affairs, Dean of School of

Medicine, University of Oklahoma Medical Center, Oklahoma City, Oklahoma

ROBERT T. DIVETT, Librarian, School of Medicine, University of New Mexico, Albuquer-que, New Mexico

ROBERT C. DUNCAN, Ph.D., Department of Biostatistics and Epidemiology, School of

Health, University of Oklahoma Medical Center, Oklahoma City, Oklahoma

LEONARD M. EDDY, Librarian, University of Oklahoma Medical Center, Oklahoma City,Oklahoma

GERALD ESCOVITZ, M.D., Acting Chief, Continuing Education Branch, Division of Phy-sician Manpower, Bureau of Health Manpower, U. S. Public Health Service, Departmentof Health, Education and Welfare, Washington, D. C.

ROBERT A. GREENES, M.D., Laboratory of Computer Science, Massachusetts General Hos-

pital, Boston, MassachusettsWILLIAM C. HARLESS, Assistant Director for Education, Computer Facility, University of

Oklahoma Medical Center, Oklahoma City, OklahomaJ. HERBERT HOLLOMON, Ph.D., President-Designate, University of Oklahoma, Norman,

OklahomaEUGENE D. JACOBSON, M.D., Chairman, Department of Physiology, University of Okla-

homa Medical Center, Oklahoma City, OklahomaHILLIARD JASON, M.D., Ed.D., Director, Office of Medical Education Research and De-

velopment, College of Human Medicine, Michigan State University, East Lansing,Michigan

ALAN 1ZAPLAN, MD., Acting Assistant Chief, Physician Education Branch, Division ofPhysician Manpower, Bureau of Health Manpower, U. S. Public Health Service, Depart-ment of Health, Education and Welfare, Washington, D. C.

DONALD A. B. LINDBERG, M.D., Director, Medical Center Computer Program, Universityof Missouri, Columbia, Missouri.

THOMAS N. LYNN, M.D., Vice Chairman, Preventive Medicine and Public Health, Univer-sity of Oklahoma Medical Center, Oklahoma City, Oklahoma.

FRANK W. McKEE, M.D., Directcrt, Division of Physician Manpower, Bureau of HealthManpower, U. S. Public Health Service, Department of Health, Education and Welfare,Washington, D. C.

GEORGE E. MILLER, M.D., Director, Office of Research in Medical Education, College ofMedicine, University of Illinois, Chicago, Illinois

TAMES G. MILLER, M.D., Ph.D., Principal Scientist, EDUCOM, and Academic Vice-Presi-dent, Cleveland State University, Cleveland, Ohio

ARTHUR W. NUNNERY, M.D., Department of Pediatrics, University of Oklahoma MedicalCenter, Oklahoma City, Oklahoma

THEODORE PETERSON, PhD., Computer-Aided Instruction Laboratory, Harvard Univer-sity, Cambridge, Mamachusetts

IRWIN PIZER, Librarian, Upstate Medical Center, State University of New York, Syracuse,New York

9

ALEXANDER SCHMIDT, M.D., Chief, Continuing Education and Training Branch, Divisionof Regional Medical Programs, National Institutes of Health, Department of Health,Education and Welfare, Bethesda, Maryland

MITCHELL SCHOROW, Chief Instructional Systems Section, College q Medicine, Univer-sity of Illinois, Chicago, Illinois

LEE S. SHULMAN, Ph.D., Associate Professor of Education Psychology and Medical Edu-cation, College of Human Medicine. Michigan State University, East Lansing, Michigan

JOHN A. STARKWEATHER, Ph.D., Coordinator of Information System; University of Cali-fornia San Francisco Medical Center, San Francisco, California

JAMES SWEENEY, Ph.D., Director, Computer Laboratories, University of Oklahoma, Nor-man, Oklahoma

NORMAN E. TUCKER, Education Consultant, Continuing Education Branch, Division ofPhysician Manpower, U. S. Public Health Service, Department of Health Education andWelfare, Washington, D. C.

JOSEPH M. WHITE, M.D., Dean of the Medical Faculty, University of Oklahoma School ofMedicine, Oklahoma City, Oklahoma

STEWART WOLF, M.D., Regents Professor of Medicine, University of Oklahoma MedicalCenter, Oklahoma City, Oklahoma

10

9

MEDICAL TECHNOLOGY AND SOCIAL CHANGE

J. HERBERT HOLLOMON, PH.D.'

Good morning, ladies and gentlemen.Welcome to the conference on computerand medical technology.

I speak to you on a subject on which Iam not an expert. So I would like tomake the subject somewhat more gener-al at first. I wish to discuss the generalrelationship between technology and so-cial change. We face a time in history inwhich many people believe that technol-ogy and its consequences have becomeso harmful to the society that this harmoverrides the possible good that technol-ogy brings. People young people par-ticularin perhaps more perceptive, morecaring, more affluent than any youngpeople of any society have begun toask questions: "How is it that you allowtechnology to pollute the atmosphere,when you build factories? How is it thatthe people who utilize technology allowthe streams to be polluted so that ourwater supplies are undrinkable?" "Howis it that a society can allow a system tooperate that allows 50,000 people to bekilled in highly technical devices calledautomobiles, and 2 or 3 or 4 million,whatever the appropriate number,wounded?" "How can a society not con-sider deeply the problems that are creat-ed with the introduction. of agriculturaltechnology which so improves the lot ofthe farmer and the efficiency of the ag-ricultural process that it displaces themarginal worker and forces him into thecity and leaves him there to sulk and tobecome discouraged and become dis-connected with the very society whichthat technology is supposed to server'Mow can a society be so inhumane asto bring new medical technology to a

country to decrease the birth rate, in-crease the life expectancy and decreasethe death rate and allow those people tostarve?" The young person continues toask, "What is the value of a system thatdoes not consider slow starvation asdeadly and as devastating as a suddendeath?" "What," says the young fellow,

"is the problem with a society that doesnot seem to be able to control the natureof the weaponry that technology brings

whether that weaponry is nuclearbombs or chemical and biological war-fare?" That's one side of the question.

And the other side of the question isthat the devices of technology, the engi-

neering achievements of man over thou-sands of years, thousands of centuries inhis evolution, have made it possible forman to survive at all. It seems almostabundantly clear that without weapons,man could not have survived and devel-oped in an evolutionary way the deepculture and the nature of the society.The weapon became a tool, and the toolmade it possible to remove much of theburden, at least of the affluent society,from man's back. It extended his arms; itplowed his fields; it helped in the treat-ment of disease. Technology has beenthe means by which it has become possi-ble for man to use his creative talents to

do something more than be a slave tothe driving necessity of having food andshelter and clothing and protection.

Technology has two sides. It has onthe one side the immediate and directbenefits that it brings in the treatment,the caring, the feeding, the housing, andthe transportation of people. And it hasanother side, and that is its potential

Vresident-Designate, The University of Oklahoma, Norman, Oklahoma

11

harmful effects on the society in general.This dichotomy, this difference in view-point, causes the deepest possible divi-sion in those people who are of goodheart as to what is the proper use oftechnology for the benefit of society,and who is to decide? It strikes mestrange, that in engineering, in medi-cine, and science, the people who areconcerned with the application of tech-nology to the useful pursuits of man, arenot at the same time deeply concernedwith the ethical, moral, political and le-gal questions which that very technolo-gy produces in its introduction. Let meask you a few questions: Who shall de-cide when and under what conditionsartificial devices will be used to extendman's life? Shall we leave it to doctors,as presently trained, doctors who arehighly capable and highly skilled techni-cal people, but who are largely notknowledgeable of the law, of the socie-ty, of religion, faith, values and ethics?Having had to spend so much time be-coming the doctor that he is, is it possi-ble for him to decide the deepest kind ofvalue judgments alone, such as whoshall benefit from the technology whichis just around the corner? Is it proper,for example, for a decision to be madeas to which of two patients gets an arti-ficial kidney because one of them hap-pens to have more life insurance thanthe other and therefore his family canbe better cared for? I wonder.

The technology which man has devel-oped has made it possible for the socialchange in every modern society. Tech-nology and its development led to theindustrial revolution. Technology and itsdevelopment led to the development ofthe modern city. Technology and its de-velopment made it possible for only 6%of the people in the United States tofeed all the rest of us. Technology madeit possible to handle information in suchlarge quantities that no man can possi-

12

bly compete with it as far as the storage,retrieval and codification of that infor-mation. Technology will, in fact, in thefuture make it possible for man tochange the genetic character of menthemselves. The technology that createdthe drugs so wonderful in the treatmentof disease, is yet the very beginning of atechnology which will so infuse educa-tion and medicine that we can speak ofthe present time, not as an industrialrevolution, but as an information revolu-tion. We are at the stage in the societywhere the extension of man's arm bymachines and tools is now nearly com-plete, but the extension of man's mindand perception in the seeng, evaluating,handling, storing, and projecting infor-mation has only begun. And if you thinktoday's sophisticated computers whichare able to program themselves are so-phisticated, if you think that informationcensoring devices that can be used re-motely are sophisticated, I will remindyou that they are yet but the very sim-ple and nascent beginning of what willbe, for we have now discovered, like thediscovery of steam power and the me-chanics of converting heat to energy,that it is possible to manipulate, controland calculate the handling of informa-tion. And we have discovered simulta-neously that it is possible to do that on avery small scale, on a scale that does notyet reach the scale of the neural net-works of the brain, but on a scale whichin a decade may be so. We will be ableas well to produce the kind of connec-tions in a computer, that will make itpossible for one to say truly that a ma-chine thinks, rather than just solves

problems.All of these things raise the deepest

kinds of questions questions, first, asto how these techniques are to be usedand to what end, and how they shall af-fect the very people who use them, andwhat resistence will the practitioners of

the old arts put up, for it is also truethat technology, by being the mother ofsocial change in many regards, is alsoresisted deeply and often effectively bythe very people who have to put thetechnology to use. I was reminded oneday of a conversation with PresidentJohn Kennedy and one of his economicsadvisors. We were trying to persuadehim and the President that everythingthat was being done except the finalevaluation of the state of the ecohomyby the economists could be done betterby computer, and that we were not us-ing computers to model the economy,but rather we were using some sort ofsubjective ad hoc judgment. I'll neverforget the day we tried to persuadethem that technology would make wiserjudgment possible by economists if theywould but use it to model the economy.The Economist said to me, "Herb, butthen what would I do?" And you know,that's sort of an odd reaction a reac-tion, however, that is completely under-standable, for all of us depend for oursecurity on the things we've learnedwhen young, and we depend for ourjobs on the techniques and tools of adecade or two ago.

Technology, specifically in medicine,will lead clearly to all sorts of artificialdevices and implants. And here thequestion comes, just as in the case ofdrugs, under which conditions shall theybe used, what tests shall be requiredand what judgment shall be made?These questions will be applied, evenmore specifically to computers, and Idon't think the computer is nearly as im-portant as the whole process of handlinginformation. It is the whole system. It is

the conversion, the calculating, the scor-ing, the retrieving, the displaying, andthe orderly logical sequence of the proc-ess. All too often, computers are usedsimply to replace manual operations.The computer provides, however, atechnique and a possibility of doingthings with information that man can'tdo himself at all. And that's why it is soimportant for us to consider how com-puters and information systems will infact affect diagnostic and clinical aid,work on organizational or administrativeproblems, and most importantly, aideducational tools. But they will not workif they are a mere substitute for whathumans are now doing, and they will notwork unless the computer and the infor-mation system is deeply tied to the verypeople who have to use the informationonce it is produced. Information systemsin medicine will make it possible to aidlarge numbers of people with relativelyfewer professionally trained medicalpeople. That doesn't mean we won'tneed more trained medical people, butthe relative number should be smaller.

One of the things that is so importantin this country is that we have commit-ted ourselves to two deep responsibili-ties about people. One is, that they shallhave the opportunity to have all the ed-ucation that their ability and motivationallows them, independently of theirwealth, sex, and color. This ideal is verycostly. The second ideal is that no manshould be without adequate medicalservices because of his position, his race,his color. If we are to do these twothings at a price we can pay, our onlyhope is to use technology in its mostbeneficial way.

OBJECTIVES OF CONFERENCE

JOSEPH M. WHITE, M.D.

Mr. Chairman, Ladies, and Gentlemen:I wish to add my welcome to that of Dr. Dennis and Dr. Hollomon. I sin-

cerely hope that you will enjoy your visit with us and that you will find theconference both interesting and stimulating. Our increasing capabilities to dis-cover new uses for the modem computer plus the enormous efforts now beingexpended in almost all our medical schools in modification and revision of themedical curriculum make this conference a timely one.

The use of computers as an educational tool is both direct and indirect.Indirect applications include the support of the administrative and researchactivities of a Medical School and, more recently, the processing of patientcare information. Direct applications to medical education are now beginning to

appear. These include information retrieval and the use of the computer as aninstructional tool. It is this latter role that we will be concerned with in detailduring this conference. However, we must keep the other educational applica-tions in mind and fully explore the potential for them.

In any instnictional system, three general problem areas must be considered.

These have been described as the technical, semantic and effectiveness prob-lems. The technical problem is concerned with the hardware and equipmentcapabilities required to implement the system. The semantic problem has to dowith the relationship between the meaning intended by the instructor and theinterpretation made 'by the student. That is, design of a system which will ac-curately convey the instructor's thoughts to the student. Finally, the effectivenessof an instructional system is concerned with those changes induced in the student

by the instructional experience. Each of these should be carefully explored be-fore introduction of any new system.

Before discussing the specific objectives of this conference, it would be well

to mention some of the thoughts of others regarding the role of computers inan educational setting. Three major roles have been defined:

1. to serve as a resource to the student for study;

2. to assist the instructor in his own study and preparation and in provisionof timely summaries of student progress individually and as a group; and

3. to provide assistance to the learning researcher. Some experience has beengained in each of these areas, especially in the undergraduate collegecurricula.

The following general questions must be explored if the objectives of thisconference are to be realized-

1. What is now being done with computers in Medical Education?

2. What are the realistic potentials of computers in Medical Education?

3. What type of work must be done to realize that potential and how shouldthis proceed?

*Dean, Medical Faculty, University of Oklalwma Medical Center

14

These general objectives invoke a number of more specific considerations

including:What are the technical problems that must be overcome? What constitutes

an effective system?Perhaps more importantly, what requirements are placed on the faculty and

students to enable them to make effective use of the computer based system?

How can training programs be developed to permit them to meet those require-

ments?Will students and physicians accept the direct use of computers as a part of

their educational process? More specifically, will they utilize the potential of the

computer by sitting at a terminal and interacting with an instructional program?

How do we get them to take advantage of it?

How can the total effectiveness of this means of instruction be evaluated

and, obviously, it must!How should such a computer based instructional system fit into a modern

curriculum? Should it serve primarily as a learning resource for the student or

can it in fact be a basic part of his primaly instruction? Should such a system be

independent or should it be a part of a larger information system?

Although it is not possible at this time to provide the specific answers to all

of these questions, we are hopeful that this conference can generate ideas and

directions to follow if we are to find the answers.

The conference has been structured so that each of you will be able to par-

ticipate actively in the discussion and as educators your ideas are needed. We

feel that the more or less informal nature of this conference will permit an envi-

ronment leading to full opportunity for you to express your own opinions and

ideas. If this conference is to be a success it will mean that all sides of the issue

have been explored. We are hopeful that this will be done.

Thank you very much.

15

att.,V1,4

STATEMENT OF THE SURGEON GENERAL*

WILLIAM H. STEWART, M.D.

Medical education is challenged today as never before in its history.The needs and aspirations fos medical care expressed by the Americanpeople demand rapid quantitative growth. Our professional commit-ment requires that this growth must not be accomplished at the ex-pense of quality. The explosive development of medical science andtechnology, by raising the potential for excellence ever higher, makesthe production of such excellence increasingly complex and difficult.

For these reasons it is vitally important that medical education drawupon the best in educational technology. We simply cannot meet thenew challenge with the old familiar methods. Yet, at the same time, itis equally imperative that we be selective in adopting and adaptingnew methodology. We cannot afford to accept new methods just be-cause they .are new, or abandon old ones because they are old. Weneed to measure such means against our fundamental purposes andchoose accordingly.

Therefore, this conference is not mei* timely; it is of the highesturgency. In wishing you a most productive and successful meeting, Iam wishing more than that: that this conference may represent asignificant stride toward more productive and successful medical

ucation in the years ahead.

Transmitted to Conference on the Use of Computers in Medical Edacation by Frank W. McKee,M.D., at opening plenary session.

EXCELLENCE -

FRANK W.

It is always a pleasure and privilegeto participate in meetings dedicated tothe study and improvement of medicaleducation, and the exploration of betterand more efficient ways of coping with

this fundamental responsibility. I amparticularly honored to be invited to

speak before you begin your work inpursuit of the "Objectives of the Confer-ence" that have just been laid before

you, because this offers me the opportu-nity to share with you somc broader per-spectives of medical education intowhich these immediate objectives andyour activities in support of them mayfit. I should point out the rather obvious

fact that my position on the programshould convey to you that I am not in.

any sense an expert in the use of com-puters as an educational tool, or will Ipresent any new experimental or otherdata about the subject of the confer-ence; otherwise I would be includedwith the quality cast you have assem-bled for that purpose. Therefore, I hap-pily accept my role as a layer of founda-tions, and will watch with interest theemerging structure of new contributionsto the medical education process whichyou erect from man-made tools and sys-tems.

As many of you know, I am privilegedto head the Division of Physician Man-power in the Bureau of Health Man-power. This Division is the focal pointfor medical education in the Public

BUY OR BEGET?

McKEE, M.D.'

Health Service. Within the responsibili-ties of our assignment are those relatedto the major support of educational pro-grams, both in medical school curriculaand in continuing education activities, inthe construction of new schools for allthe health professions, additions and ren-ovations to existing schools, and in sup-port of hospital and patient care facili-ties which are intimately concerned withthe educational process. The newly pro-grammed Basic and Special Improve-ment Grants, which are the Federal gov-ernment's first clearly defined funds for:;upport of educational operations, arewithin our review mechanisms. Throughresearch grants and contracts, support isoffered for a variety of investigations of

new approaches to educational format,with the pious hope of a dividend of ac-complishing some progress in the educa-tion of both medical students and prac-ticing physicians while the experimentsgo on. All of these efforts are compo-nents of the overall mission of the Bu-reau of Health Manpower which is di-rected towards a quantitative and quali-tative improvement in all the healthprofessions in the provision of medicalcare for all the American people. In anutshell, we are interested in increasingthe number of qualified individuals whowill responsibly take care of the sick,and act effectively in disease preventionand control, notwithstanding the otherimportant considerations for develop-

Director, Division of Physician Manpower, Bureau of Health Manpower PHS, DHEW

17

ment of new knowledge and the on-going study of disease processes.

Having outlined, then, the specific ob-jective of increased numbers of qualifiedmanpower, and having described the in-terests of my particular Division in thephysician portion of this goal, and thevarious ways in which the production ofphysician manpower is supported, it isappropriate to offer some observationsabout the educative process itself and itsdominant and pervasive role in theachievement of these national objec-tives. The following comments are notbased on second hand or hearsay infor-mation, for, in dealing with all of thecolleges of medicine and osteopathy,and having the privilege of assisting ournational legislators and others in re-plying to a wide spectrum of inquiryand criticism of present medical caresystems in the United States, one devel-ops a reasonably accurate, continuing,and rather acute familiarity with what avariety of our citizens think is wrong orlacking. In our ability and agility to re-spond to such inquiries, a number ofproblems are further identified and pos-sible solutions are generated.

The concerns of the public, as pre-sented in correspondence, and in otherinquiry, fall into two categories. Thefirst is that there are not enough physi-cians doirg the right things in the rightplaces. Rural areas, and inner citiesseem to be the most easily identified des-erts of health care, but other geograph-ic areas and even our utopia of modemliving, the suburb, has its problems. Thefourth branch of.the Division of Physi-cian Manpower, to which I have notheretofore alluded, Physician Utilizationand Supply, is concerned with solutionsto the problems of physician distribu-tion, through encouragement of studiesof more effective practice patterns,investigations of why major gaps inhealth personnel exist, and the means

18

that might be best used to correct theseunfortunate, and occasionally desperate,situations.

The second major concern of the pub-lic relates to the inability of large num-bers of apparently eminently qualifiedyoung men and women to gain admis-sion to medical school. Letters on thissubject are directed to a number of gov-ernment officials with pleas for interces-sion on behalf of this or that individual,a practice which is consistently regardedas meddling in the affairs of a school orschools. But, be that as it may, whenthere is public clamor, there will be ac-tion, and it is our hope that the expan-sion of established schools and the con-struction of new schools will provideincreased educational opportunity forevery qualified applicant in the UnitedStates.

These two citations are foci of publicattention and represent a surfacing ofdeeper and more complicated concernsabout medical education which are inthe purview of all of us who are in-volved both in providing it and support-ing it. These latter problems are many,with a large overlay of human nature inthem, which often, surprisingly, is moreserious than financial problems. Of late,a number of cliches to identify elementsof these problems have crept into thelanguage of medical educators, andwhile some of these are tiresome andirksome and subject to interpretation bythe one using the term or phrase, theydo bespeak both complacency and, par-adoxically, a certain insecurity and"whistling in the dark to keep up cour-age" sort of attitude. Since I hardly havetime today to run through a glossary ofterms used by medical and other educa-tors to describe the ideals, the goals, theproblems, and the underlying financialneeds, yea, even demands, for theirachievement or solution, I will mentiononly one "EXCELLENCE."

Excellence is about as complete aword to define educational aims as onecould wish, and has as synonyms su-

periority, preeminence, worth, value,goodness, purity, and greatness. Its es-sential meaning is merit and virtue, andcould well join with motherhood, peace,the American flag, and the reduction oftaxes as a standard under which wecould all join. Excellence appears with

an element of glibness in medical schoolpublications, particularly those directedtowards alumni and granting agencies,and I quote one such:

"In the face of pressures for change, itis especially important that we keep ourprimary goals and responsibilities clear-

ly before us."The first of these is excellence in ed-

ucation. This arises from the capabilitiesof the staff, the quality of the studentbody, and the continuous upgrading of

our educational program. It arises, too,from the commitment of the staff to re-cognizing the individual qualities ofeach student and building upon them.Although we will not invariably be suc-cessful, we strive for an environment inwhich understanding and enjoyment of

education are shared by students andstaff alike. As pressures rise for the edu-cation of more physicians, we must ex-amine our educational goals, explore in-novations, and establish our programson the basis of the best evidence avail-able."

This is what I would call a completestatement, and while an expression ofbut one school, it could well serve all,

and be a preamble to whatever mightfollow as definitions of accomplishmentor requests from any and every sourcefor support for faculty, students, and fa-cilities. It is hard to fault as an expres-sion of goals and implies all of the defi-nitions and syncnyms of excellence in itsstated quest. But while it can escapecriticism as an ideal, it most certainly, to

19

those who know about processes of im-plementation, lays itself bare to searchingquestion, and the "feet of clay" on thequicksand of its major premise. In pur-suit of excellence, medical education hasassumed a flavor of Madison Avenueand the supermarket, with "capabilitiesof the faculty" prominently mentionedas a pillar in generating "excellence ineducation."

Perhaps these "capabilities" should bescrutinized more closely to determinewhat they are, what they should be, andwhether they are actually directed atthe purposes of medical education andits ultimate goal of helping solve theproblems of a real world of people, dis-ease, and untimely death.

Some years ago, before the enormou.sinvestment of the Federal Governmentin support of medical research, medicalschools followed integrated programsembodying teaching, research, and pa-tient care with minimum categorization,because of the recognition of the impor-tance of all these elements in physicianeducation. In recent times, however, thiscategorization has been carefully identi-fied, the components have been legisla-tively isolated, and the integrity of theschool as a common center for medicaleducation has been greatly changed.This situation has led, in turn, to con-siderable change in the faculty, thosewho constitute it, and what the ideas ofboth the school and the faculty memberare about correlated responsibility. Re-cruitment of faculty is now a highlycompetitive exercise for trained individ-uals in short supply and is characterizedby high interest and investigation of al-most everything about the candidatethat has to do with prestige, funds, andpersonal contribution, in varying mean-ings of the term. Strange and unwar-ranted assumptions about teaching qual-ifications, character as a man, and as amodel fct students, are often made, and

may be covered by a statement such as,"He s a nice guy, and the students seemto like him." The candidate himself isusually more captivated by the personaladvantage of the situation, and concernshimself with his space, his arena of con-trol, be it medical specialty or research,his teaching "load,' not opportunity,"and what colleagues he will find worthyto help continue the rise of his star.From this sort of activity, then, comesone aspect of faculty capability, whichis hardly conducive to a program ofmedical education directing young menand women to preparation for assumingmajor responsibilities in the care of sickpeople.

A second factor in faculty capabilityis the actual involvement of patient care.At one time, all medical schools hadnumbers of competent full-time seriousclinicians, who not only taught medi-cine, but personally practiced it. Theywere backed up by a substantial part-time or volunteer faculty who were ac-tually engaged in private medical prac-tice, and brought to the students andhouse officers a true flavor of day-to-day practice to complement other lessonsof greater breadth or depth. With thegrowing ability, and incidentally, realneed to attract funds through the mech-anism of research and its considerablesupport, more and more full-time facultywere added to the medical schools. Con-comitantly, burdens of increasingly de-manding practice greatly diminished thetime available to the private practitionerto participate in regular teaching sched-ules. But w'iatever the factors responsi-ble, this chal tge in personnel has done agrave disservice to many students whohave a conscientious desire to becomeaccomplished practitioners outside auniversity or medical school setting, andhas effected a considerable diminutionin "capability of the faculty" in itsbroader sense. Some expanded use of af-

20

13

Mated hospitals has compensated to adegree for this change, and perhaps ex-plains the popularity of these programs,a popularity that often mystifies thefull-time faculty member in his universi-t)., hospital office.

A third factor in determining "capa-bility of the faculty" is whether there isany general understanding of the rea-sons for the existence of a particularmedical school, and what its major inter-nal role might be, as well as its responsi-bility in the local, state, or national com-plex. Such understanding should involvethe President and the Board of Trusteesas well as the Dean and the faculty. Insome instances, where a single schoolexists in a given state, this responsibilityshould be quite clear, but I have beenamazed in a recent visit to one suchstate to find the wide differences ofopinion of the state legislators and themedical school administration as to thefundamental purposes of the medicalschool. However, if the true goals of theschool are defined, and generally agreedupon, and if there is the intention toproduce qualified graduates, who willcontribute to the real world of dealingwith illness and its prevention and con-trol, and who will be educated to devel-op high standards of responsibility andhuman concern, then the "capability ofthe faculty" is more easily attested, andthe achievement of excellence is a realpossibility. How strange it seems, in thisday of heart transplants, Nobel and oth-er prestige prizes, and enormous indi-vidual project grants, that excellencecan be pursued as a goal within itself. Itactually can be more closely approxi-mated by accepting a clear and obviousresponsibility, meeting it, and findingsatisfaction in a job well done.

Excellence, true excellence, was oncedeveloped and nurtured as a home-grown product and expressed itself in adedication and loyalty to the institution

and understanding of its purposes which

was truly representative of "capabilityof the faculty." The teamwork was theresult of years of associations with es-tablished trust and mutual interest. The

final product graduated physiciansreflected a stability, responsibility, andconcern which was true excellence. TheWorld Series will always be a better dis-play of baseball than.the All-Star games,and the school that begets its own excel-

lence will always have more to offer itsstudents and the people for whom it isresponsible, than purchased "excel-

lence."From this inherent capability, then,

comes the greater wisdom for designinga dynamic, competent, and inspiring ed-ucation program, which might really beaccepted by the faculty responsible forit. The "quality of the student body"cannot fail to be enhanced, with per-haps some resolution of problems of ad-mission and the alarming attrition of ad-mitted students, and some true studententhusiasm for "the continuous (ly) up-graded educational program." Concernfor direct patient care, coupled withgood educational techniques, will amel-iorate public anxiety about physiciansupply and availability. Such concernwill also have a salutary effect on thewasteful and distressing problem of in-creasing student attrition.

A profession as essentially human asmedicine must have a firm base in thecharacter of its primary seat of produc-tion the medical school, for it is fromthis focus that the true quality of profes-sional excellence emanates.

On behalf of this ideal, it is the policyof this Bureau and Division to supportinstitutions and strengthen them

through their chosen administrative

2

21

channels, so that the determination ofthe essential educational need is neitherdissipated nor eroded by less importantconsiderations. It is equally importantfor the Association of American MedicalColleges, acting through its individualconstituents, to review the whole spec-trum of faculty acquisition and to en-courage principles and procedures moreconsonant with the needs and purposesof the system.

The role of this conference is to dis-

cuss among knowledgeable people anexciting adjunct to the processes of med-ical education, which can offer real as-sistance in solving problems of increas-ing information and demands for accel-erated production. I look forward tosome interesting revelations and accom-plishments, but the keystone of this key-note address is that the success or fail-ure of our system of medical education,in its multitude of individually tailoredand variously directed programs, de-pends on the character and the qualityof its leaders and proponents. The re-finement of procedure, the innovation ofmethod, the desire for excellence as ex-pressed through graduates dedicated tomedical care in all its facets, will alwaysrely for their success upon the humanbeings who choose to accept responsibil-ity for the achievement of the goals ofthe profession, not the particular whimof a certain school or individual. Thegoals of the profession were definedlong ago by Hippocrates, Jesus of Naza-reth, Maimonides, and have been ear-nestly supported through the ages byconsiderable numbers of men of dedica-tion and principle. Perhaps we shouldall devote more attention to this funda-mental .matter, and I would urge yourpersonal consideration.

7CW..7. [NWT/ .9.2,111.

EVALUATION OF THE USE OF COWLMEDICAL EDUCATION

STEPHEN ABRAHAMSON, Ph.D.°

The use of emerging computer tech-nology in education is, of course, quiterecent. And, as is the case of the appli-cation of educational innovation to med-ical education, the introduction of com-puters as an aid to medical educationcomes significantly later. Thus, whenone asks questions concerning evalua-tion of the use of computers, answerscannot really be found in data at thistime. At least, data in substantial and/orsignificant amounts do not exist. Thus,this paper cannot assume the familiarform of a review of research or of a syn-thesis of published (or unpublished ) re-ports. Rather, at this stage of develop-ment of the application of computertechnology to medical education, onecan only look forward to the design ofstudies and collection of data.

Consequently, exercising familiar li-cense, the author is changing the topic.The modification truly reflects thechronological sequence of events in edu-cational innovation, incidentally, and at-tempts to warn through some discus-sion of the days ahead. What factorsmust be considered as evaluation is

planned and experiments are conducted?What data are needed to help us makeinformed judgments? What kind of re-search design will provide an adequatetest of the use of computers in medicaleducation?

Despite the fact that millions of wordsof glittering prose ( and possibly some of

poetry ) have been wof evaluation, some rethe outset so that lanot become snaggedof conflicting definitimutual understandingnot be achieved beuin just plain definition

Avoiding the useand an extended pre .

mal definition ( evaltuous process, based tcratively developed, athe status of and chbehavior ) one mustthe steps in evaluaticmany times it has bematter how many diflwritten these things, (sists of these componc

(1) Evaluation stildefinition of obfectivhas been made so oftpeople that listenaweary of hearing (,dictum. But it is nornonetheless importan.

unfortunately diseems that becauseabout objectives" forhave somehow thrc

rid ourselves of thejectives, 'Such, is far figard to any educatioercise the questions nanswered): What sl

°Director, Division of Research in Medical Education, School of Medicine,ern California, Los Angeles, California

22

21--

know at the end of this educational ex-perience? What should they know howto do? How else should they bechanged?

(2) Evaluation still demands criteriaby which to judge students' achieve-ments of those objectives. While theseare sometimes embodied in a good state-ment of objectives, more frequently theyare omitted. How much new informa-tion should students have acquired?How proficient should they have be-come in the execution of specific skills?These questions should be answered, in-cidentally, before the attempt to meas-ure achievement and not afterward byscrutinizing scores of other records andassigning some "logical" (statistical orotherwise) cutoff point or points.

(3) Evaluation still requires measure-ment of the achievement of each of theobjectives. And educational measure-ment is best accomplished uni-dimen-sionally. That is, acquisition of informa-tion is most appropriately and reliablymeasured by application of someachievement-test technique pencil-and-paper, perhaps. The measurement ofachievement of slcills necessitates someobservation of the students in perform-ance of those skills either directly orthrough the aid of electronics in theform of video-tape. But measurement, inwhatever form, provides the informationconcerning student progress needed forthe fourth step.

(4) Evaluation still includes imple-mentation in the form of action involv-ing the student, the program, the objec-tives or even the profession. One muststudy the results of measurement in thelight of stated objectives, and accordingto the criteria make decisions and takeaction. Here is the judgmental phase ofevaluation. But it should not stop shortof the accompanying action to revise theprogram, review the objectives, certifythe students, or initiate new studies of

23

students, program, or profession.Concerns in evaluation of the use of

computers in medical education are per-haps best expressed in these four ques-tions.1. Does the use of computers facilitate

learning?2. If so, does it do so at the expense of

other objectives?3. What is the cost of the use of com-

puters to facilitate learning?4. Does the use of computers result in

less faculty and/or student input oftime and/or effort?

Essentially three areas of study are in-volved: (1) effectiveness, (2) cost, and(3) efficiency. The first two questionsabove deal with effectiveness, while thethird and fourth refer to cost and effi-ciency, respectively. All of these becomemajor concerns in any educational ven-ture but particularly in the situationinvolving the introduction of innovation.There is no quarrel with the contentionthat conventional and standard and tra-ditional methods in education should bestudied critically. For all too long edu-cation has enjoyed the luxury of as-sumption of achievement rather thanhaving to test its processes critically. Ithas been said that physicians have suchan enviable record of achievement inapplication of their professional per-formance perhaps because half their pa-tients would get better anyway. It is un-doubtedly true also, in that vein, thatteachers (and traditional education)can boast of having achieved instruc-tional .goals because students wouldlearn anyway!! Thus, all educationshould be, subject to careful study. Butthe real burden is upon the innovator, aswe all know. The introduction of changein teaching methods must be accompa-nied by careful study. Especially as weoperate in areas involving increasedcosts or potentially increased costs is

it incumbent upon us to document our

success, or lack thereof. There is onecomforting thought, incidentally. Themore we attempt to study impact of ed-ucational innovation, the more we man-age to learn about conventional educa-tional processes. The more we applyourselves to tasks of educational mea-surement with regard to studying out-comes of newer methods, the more re-fined our measurement techniques be-come and these can then be appliedin any educational milieu. The messagereally is to avoid any sense of paranoia,to avoid a posture of "why me?" whenasked to justify in educational out-come terms the use of computers inmedical education.

Effectiveness Does the use of com-puters facilitate learning? If so, does itdo so at the expense of achievement ofother educational objectives? The studyof effectiveness cannot even begin, letalone proceed, without a clear statementof objectives. One cannot determinewhether learning has been facilitatedwithout the answer to a prior question:learning of what? At the same time, onemust keep in mind that there are usuallymany objectives being sought al-though perhaps not all in that specificapplication of instructional media. Nev-ertheless, they must be considered. Forinstance, if a technique is employed tofacilitate the acquisition of information,it is important to know, of course,whether it did so. It is equally impor-tant, however, to discover whether theuse of that technique also ( accidentally,inadvertently, incidentally) negativelyinfluenced student attitudes towardlearning, toward continued study, oreven toward particular aspects of pa-tient care. It is possible for one learningexperience to promote achievement ofone set of objectives while significantlydeterring achievement of another. Stud-ies of effectiveness should try to includesome measurement of achievement of

24

MM 1,11S,/,

other objectives, hopefully representa-tive of all of the educational objectivesof the total educational effort. In asense, rival hypotheses must be statedand tested.

In studying effectiveness, investiga-tors must also avoid the error of testingfor educational gain only at the comple-tion of the instructional process. Howlong do students retain the new informa-tion? Over how long a period of timecan students still perform the skills theylearned? Data collected in other studiessuggest rather strongly that retentionmay be related to certain other varia-bles: utility, motivation, experience.Studies of retention of information andskills gained through the use of comput-ers in medical education, thus, will helpin the understanding of this aspect ofeducation.

Cost It seems a shame to have tospend time on the question of cost.Here, again, is a question not reallyraised with regard to many other modesof instruction. Unfortunately, the bur-den of proof is on the innovator. Andthe initial cost of equipment plus the ex-pense of operation and maintenance arenot insignificant. In addition, there isthe area of preparation of "software"the subject matter preparation, the pro-grams themselves. The cost in dollarsand cents must be studied and includedin evaluation.

But that's not all the cost that must bestudied. In addition, the input of facultytime and effort in monitoring, in evalu-ating student progress, in just learninghow to prepare programs and use thecomputer hardware must be considered.Finally, revision and modification costsmust also be calculated. Such costs inconventional education are minimal andfrankly not considered. Where the tech-nology demands efforts beyond the nor-mal range, a systematic examination isin order.

Efficiency Does the use of comput-ers affect faculty time and effort neces-sary to achieve the educational goals?By the same token, what is the effect onstudent input of time and effort? Essen-tially, a ratio of "output" to "input" is in-volved and should be studied. The studyof output must include such factors asnumber of students educated, how muchthey learned, how long they retain whatthey learned, how well they apply whatthey learned, how much they are stimu-lated to study more, and the like. Thestudy of input includes such factors ashow much time the instructors invested,how much time the students devoted tothe learning process, how many facultymembers were involved, how long ittook the students to reach required per-formance levels, and the like. It is possi-ble that an "efficiency ratio" or an "effi-ciency index" must be developed andapplied. Without it, statistical designmust be employed to allow this area tobe explored intelligently, for it is thearea in which many claims have beenmade and will continue to be made.

Another interesting area of concernmust be exposed and explored beforethis discussion moves on to suggestedstudy design. This is the area of concernfor patient safety and comfort. One per-sistent educational problem has alwaysbeen how to have beginning studentslearn skills, the performance of whichposes a threat to the safety and/or com-fort of the human object, the patient. Atthe University of Southern California,this concern led to the development of acomputer - controlled anthr op om e tricmanikin a patient simulator, whichhas come to be known as "Sim One."The original purpose was to provide asimulated patient on which beginninganesthesiology residents might .learn theskills of endotracheal intubation prior totheir having to perform the intubationof real, live patients. At least, the moti-

25

vation to begin this interesting work layin this area of concern. The stated pur-pose of the project was to test the feasi-bility of simulating a living patient, ca-pable of responding in a lifelike mannerto things done to it. In addition, effec-tiveness of its use in training anesthe-siology residents is being studied.

Sim One is quite lifc-like in appear-ance, having a skin-colored, skin-tex-tured plastic covering in the form of areal human being. Sim One has the con-figuration of a live person from thewaist up and is in the posture of lyingon an operating table, left arm extendedand ready for intravenous injection,right arm fitted with blood pressurecuff, and chest wall adorned with astethoscope taped over the approximatelocation of the heart. Sim One breathes;has a heart beat, temporal and carotidpulse, blood pressure; opens and closeshis mouth; blinks his eyes; responds tofour intravenously administered drugsand two gases administered by mask-and-bag or by tube. The important thingto note here is that Sim One has beenused in the training of residents and ef-fectiveness tests reflect a great potentialcontribution to human safety and com-fort through having beginning studentsmake their early errors and gain theirproficiency and confidence on a com-

puter-controlled "teaching aid" ratherthan on a real patient.

The test of effectiveness involved five

pairs of residents. These were assignedat random in such a way that five resi-dents were allowed to use Sim One andfive were not. The measure of proficien-

cy was a simple one. Hospital anesthesiarecords were examined by the anesthe-siology staff and a judgment was madeabout each: Would you trust this resi-dent alone in the operating room on thebasis of performance like this? Needlessto say, names and dates were concealedfrom the rater. One criterion selected

nr.-^,^,

was the achievement of four consecutive"plus" ratings; that is, the rater wouldtrust the resident in the operating roomalone on the basis of what he reviewedin that chart. While the final analysis ofdata is not available at this moment,preliminary review indicates that ittakes the non-simulator-trained residentstwice the number of trials in the operat-ing room to reach that criterion level ofperformance when compared to the sim-ulator-trained residents. Such a differ-ence obviously represents a substantialdifference in risk to patients!!

It should be noted, incidentally, thatSim One represents a remarkably imagi-native application of computer technolo-gy to problems of medical education.Here is simulation at its most sophisti-cated; a realistic body with life-like re-sponses under computer logic. The dol-lar-cost of development has been highand the cost area in evaluation is indeednot on the positive side of the judgmen-tal ledger. Effectiveness and efficiencywill be borne out by data but the combi-nation of cost and effectiveness mightbe found formidably negative in impli-cation. However, the human cost factor

patient safety must also find its wayinto the total cost-effectiveness equa-tion. Thus, there are now four areas forstudy: effectiveness, cost, efficiency,and human safety.

How should evaluation be conducted?The best single, simple answer is a for-mal, carefully controlled, experimentalstudy. Not only will evaluative data beforthcoming, but comparisons withconventional educational methods willbe possible. In this approach, one mustattempt to control for all variables ex-cept one: the use of computer. Thenwith valid and reliable measurement ofachievement level taken before and afterthe conduct of the educational experi-ence, the investigators can documentwhether the use of computers did in fact

28

"facilitate learning." And it must bepointed out that this kind of study ismost likely to catch the attention of andconvince the skeptics if the use ofcomputers does make a difference. Sincethe men who teach in medical schoolsare themselves scientists, they constantlyseek data in support of ideas. Educationis no exception for them in this regard.Furthermore, the areas of educationalachievement studied should reflect thevalue structure of medical school facultymembers. In other words, if achieve-ment in basic science areas is important,measurement of achievement in thisarea should be included always re-membering, of course, that this is trueonly to the extent that such achievementis part of the objectives.

Unfortunately, this approach is themost difficult to achieve in evaluation.Individual differences among studentspose the first problem; faculty differ-ences, the second. No two schools arereally alike enough to allow for one tobe used as a "control" for another. Intro-ducing change of any kind also managesto change the total educational settingso that speculation can be raised con-cerning whether measured change inachievement is really only a reflection ofthe act of change itself: the so-called"halo" or "Mayo' effect. And, experi-ments within one school, conductedwith time as the control variable, oftenstimulate other kinds of change withinthe school even to the extent of draw-ing a different applicant-pool of stu-dents.

As if these difficulties were notenough, another less clear problemneeds mentioning. Each instructionalmode may in fact have different goalswhich can best be served. Thus, to at-tempt a comparison in achievementwhen different .teaching methods areused might start with a major bias if theachievement of the same educational

objectives is attempted using methodstruly adaptable to quite divergent goals.But at this stage, such a considerafionmay be less important. It is presently inthe best interest of medical educationand of the use of computers that experi-mental design be accomplished wherev-er humanly possible.

Where experimental design is not fea-sible, however, there is still the need forgood, thorough evaluation;Has the edu-cational experience really' brought aboutthe changes in students that were de-sired? (Were the objectives realized?)How much faculty time and effort wererequired? (How much did it cost?)Such an approach required "only" thedefinition of objectives, the statement ofcriteria, the selection or construction ofmeasurement devices, the performanceof measurement, and the interpretationof data collected. (Note the setting offof the word, "only"11) If this process isachieved imoperly, the invertig.htors willat least know whether objectives havebeen realized even if they cannot tellwhether the process represents some-thing "better" than another approach.

The role of the innovator is not aneasy one. Not only is he the guy withthe "ideas." Not only must he then find

27

4

ways of accomplishing his ideas. Notonly does he become the activist in in-troducing the innovation. Not only musthe seek and obtain support for his idea.He must also test the idea and be pre-pared to defend with data in theface of all the difficulties in obtainingsuch data, his innovation. The role ofthe true innovator is a demanding one.He must be imaginative and flexible inorder not to become "locked in" to oneapproach in the application of his genreof innovation. Computer use is a case-in-point. Storage and retrieval of infor-mation is but one of the many potentialapplications, yet for many this has be-come the end-all. He must promote thesearch for evaluative data despite thefact that there are no such data avail-able concerning conventional modes ofteaching. He must serve as "public rela-tions" man for the entire area of innova-

tion not as high-pressure salesman, in-cidentally, but as thoughtful, data-ladeninvestigator, raising questions and pro-viding data rather than exhorting, pro-moting, and promising Utopia throughpanacea. Here, then, is the essential con-tribution of evalnation of the use ofcomputers in medical education.

Ir27..i:S. .171.1.6nes.

EDUCOM-A PROGRESS REPORTJAMES G. MILLER, M.D., Ph.D.°

EDUCO M, the InteruniversityCommunications Council, is a proto-type endeavor, a research-oriented at-tempt to bring about educational in-novation. Let me begin by illustrat-ing, in a somewhat lighthearted andirrelevant fashion, the problems of in-novation and the need to move slowlyand cautiously when initiating newprograms.

A pretest had been carried out onthe effects of diet on the aging proc-ess. The experimental animals wereporpoises. The porpoise is consideredby many to be nearly as intelligent asman. They have large brains and are,in a sense, aquatic, humanoid basket-cases. The porpoises were given anumber of different feeding regimes.A diet of newly hatched sea gulls wasfound to slow down, almost to halt,the aging process. It appeared thatwhat amounted to immortality mightbe a possibility.

The pretests were so striking thatthe research group received a largeNIH program-project grant. Large,concrete porpoise pens were built onthe seashore for containing the exper-imental animals. Up in the hills, anaviary for hatching sea gulls wasbuilt. A truck was dispatched daily topick up the young fledglings and totake them down to the pens to feed tothe porpoises.

One day as the truck passed theCity Zoo on its regular run, an old,distinguished, sedate lion suddenlyand inexplicably bounded out of hiscage and into the path of the truck.The truck could not help but run over

him. The police soon arrived and ar-rested the truck driver for "transport-ing young gulls across staid lions forimmortal porpoises." If this irrelevantand irreverent fable does not illus-trate the dangers of innovation, noth-ing can.

EDUCOM was founded aboutthree years ago by a number of medi-cal deans and.chancellors, specifically,those at Duke, Virginia, State Univer-sity of New York, Rochester, Pitts-burgh, Michigan, Illinois, and Cali-fornia. It was founded to facilitateand accelerate the use of the new in-formation processing technologies,where desirable, throughout thehealth sciences, educational, researchand service activities of colleges anduniversities.

It soon became apparent that ED-UCOM's functions should not be re-stricted to the health fields butshould be extended to all academicdisciplines. While some of EDU-COM's present members of the Boardand officers are physicians, the major-ity are not. The organization has ex-panded rather rapidly although verylittle effort has been made to give itpublicity. An article in Science ap-pearing in October, 1966, is the onlypublished article describing the totalEDUCOM program. There have,however, been a number of publica-tions relating to various specific proj-ects.

At the moment, there are 89 univer-sities on our membership roll. Theyinclude about 250 campuses. EDU-COM is open to all colleges and uni-

°Principal Scientist, EDUCOM, and Academic V ice President, The Cleveland State University.

28

2 7

7.7%T.,:,,,ticrrnmave

versities of Canada, the UnitedStates, and Mexico and will bepleased to accept your institution intomembership if it is not a member.

Several universities in England,France, Czechoslovakia, Sweden,Holland, and Australia have indicat-ed an interest in carrying out similaractivities and in cooperating withAmerican universities. We intend toconsider potential international appli-cations as soon as some of our initialNorth American projects have givenrise to services which will be of realhelp to participating institutions. In-ternafional cooperation across oceanswill be facilitated by educational sa-tellites which may be in orbit four orfive years from now.

Dr. Jordan Baruch has recently be-come our new President. A memberof the MIT staff, he was formerly thehead of the MEDINET program ofGeneral Electric. With the transfer of

our national headquarters to Cam-bridge, Massachusetts, we have de-veloped a redefinition of our func-tion. In the words of our new presi-dent, a major EDUCOM function is

"to facilitate the extraorganizationalcommunication of the universities."The reason that the emphasis is onextraorganizational communication isbecause EDUCOM does not, exceptindirectly, intend to participate in ac-tivities which can be carried on aswell by an individual professor or anindividual university. We are interest-ed, rather, in helping institutions ofhigher education to share their re-sources and carry out joint programswhich can be more efficiently han-dled, improve quality, and/or cutcosts. We are concerned with rela-tionships not only among universitiesbut between them and other parts ofour society. We shall be concernedwith such extraorganizational activi-

29

28

ties as the regional medical programs,for example, which involve not onlyuniversities and medical schools, butalso private hospitals and privatepractitioners.

PANELS ON RESOURCEDEVELOPMENT

EDUCOM is presently involved inreorganizing its technical and re-search programs, forming a new se-ries of national interuniversity panels.In the past we have operated throughinteruniversity task forces. The firsttask force was responsible for theprograms in networking. The chair-man of this task force has been Dr.George W. Brown, who was thefounder of the Western Data Process-ing Center at UCLA and who has re-cently become Dean of the College ofAdministration at the University ofCalifornia at Irvine.

The original task forces were"bootstrap" operation s. Volunteerworkers from the participant universi-ties gave their time, prepared pro-grams and proposals, and carried outresearch activities. EDUCOM hasnow attained some financial stability,however, having received severalfoundation grants, industrial grants,grants or contracts from the federalgovernment, as well as institutionaldues. It will soon be feasible for eachof the panels to have full-time one ormore professional staff workers, ei-ther in the central office in Cam-bridge or in the Bethesda office,where some of the government sup-ported research activities are nowgoing on.

The network task force will bereorganized into the panel on inter-connected computers, which will beconcerned with the organizationaland structural interconnection ofcomputers, the hierarchical organiza-

tion of data within them, and humanaccess to their capabilities.

A second panel on resource devel-opment will be the panel on videoand films, concerned with the use ofvideo recordings, films, live produc-tion, closed circuit TV, and othersuch tools for augmenting the educa-tional process.

Third, the panel on micrographicdevelopment will be concerned withthe utilization of such techniques asphotochromic reproduction in storingmasses of document information, theinterphase between such systems andthe computer, and the use of such mi-crographic images.

The Work of a PanelA new panel goes through a stage

of group dynamics and a stage ofself-definition of their mission. It mayidentify areas of appropriate activityfor EDUCOM and it can suggest spe-cific projects for EDUCOM or somesubset 3f its members. Such a panelwill be different from a study sectionof the National Institutes of Health,for example, in that its major role isto initiate programs rather than to re-spond passively to programs suggest-ed by individual professors or by uni-versities. But it certainly can receivesuch suggestions. In a few cases suchsuggestions have eventuated in pro-grams that have been supported orsponsored by EDUCOM.

EDUCOM panels can communi-cate among themselves to establishjoint activities, since there will be in-tersections of interests among thepanels. And they can maintain tempo-ral continuity. The life span of agroup or organization may be farlonger than any individual human lifespan. This means that continuity ofprogramming is feasible even if indi-viduals rotate in and out of variouspanels.

30

29

Panels in the Applications and Utili-zation Area

The first of thele is the panel onextended education, concerned withthe continuing education of theprofessional, with the resumed educa-tion of those whose education hasbeen interrupted, and with the reedu-cation of those dissatisfied with theircareers. The panel on extended edu-cation is a reformulation of the taskforce on continuing education,chaired by Dr. George E. Miller, Di-rector of the Office of Research inMedical Education at the Universityof Illinois College of Medicine.

The second panel in the applica-tions and utilization area is the panelon libraries and data banks. The pan-el is concerned with the utilization ofstored information in the educationalprocess, with the instructional impli-cations of the fact that large bodies ofdata can be made rapidly accessible,and with the development of the li-brary function broadly defined. Thework in library communications wasformerly in the EDUCOM task forceon networks. It has become clear tous that, while there have been andwill continue to be important over-laps between networking activity andlibrary functions, they should bedealt with by separate panels becausethey are fundamentally independentfields.

The third EDUCOM panel on ap-plications and utilization is the panelon external affairs, concerned withlegislative matters, public policy, andsocial issues involved in educationalinformation processing. Its predeces-sor was the EDUCOM task force onlegal and related matters. It is co-chaired by Professor Benjamin Kap-lan of the Harvard Law School andProfessor Arthur Miller of the Univer-sity of Michigan Law School. The

t

new panel will include attorneys, spe-cialists in libraries, television, andcomputers, and educators. EDUCOMhas taken an active role in a numberof matters directly related to utiliza-tion of the new technologies in edu-cation, of whose importance manyuniversity faculties and administra-tions are unaware. For example, Con-gress is presently working on newcopyright legislation, the first to ap-pear before Congress in about 60years. The draft of the bill passed bythe House precluded the possibilityof operating legally such an inter-university computer, library or televi-sion network as EDUCOM envisions.EDUCOM universities and officers

sent statements or appeared at theSenate Committee hearings. Thesematters were also discussed with rep-resentatives of publishers and otherindustries. EDUCOM has had somerole in slowing action on the copy-right legislation, which was writtenbefore the importance and implica-tions of computers and networks be-came clear. The legislation can nowbe rewritten to permit adjustmentsfor new technological developmentsas they occur.

EDUCOM is also concerned withantitrust and negligence laws in theoperation of interuniversity and in-trauniversity networks. A nonlegalbut very important area concerns theethical use of confidential informa-tion. This is particularly important tophysicians in terms of patient records.

EDUCOM'S Plans for Networks

Although EDUCOM is concernedwith all applications of technologiesin education, it has devoted particu-lar attention to the sharing of re-sources through the development ofnetworks, planning to develop andoperate a prototype interuniversity

31

network. Networks are nothing new.Natural networks have been in opera-tion as 161.g as there have been col-leges and universities with one-wayor two-way communication betweentwo or more students and facultymembers. A conference is one form ofoperation of a natural network.

A different kind of network is aphysical network like the telephonenetwork or the mails. A thoroughanalysis of communication among

and within universities requireS acomparative evaluation of traditionaland new networks. Because a net-work employs a new technology ob-viously does not mean, of course, thatit is better.

At the same time that EDUCOM isinvolved in planning electronic net-works, it is also developing an organi-zational network. Each of our mem-ber universities has an institutionalrepresentative w h o communicateswith his campus and represents hisuniversity with a vote on the EDU-COM Council. The Council elects theBoard of Trustees and the Board ofTrustees selects the permanent offi-cers and full-time staff. If we place aprototype network into operation, itmust have its own organizationalstructure as well.

EDUNETThe most intensive planning activi-

ty related to networks in EDUCOMwas accomplished at the Universityof Colorado at Boulder in July, 1966.The conference was funded by agrant from the National Library ofMedicine, the National Institutes ofHealth, the National Science Founda-tion, and the Office of Education.About 185 people attended, repre-senting each of the 42 universitiesthen members of EDUCOM, the ma-jor long line and hardware compa-nies, publishers, approximately 15

government agencies, and a numberof foundations. The conference prod-uced a book called EDUNET, a re-port of the conference including opin-ions of the participants and a set ofrecommendations.

The conference almost unanimous-ly agreed that there were real needsin colleges and universities for an in-teruniversity network.

It was unanimously agreed thatmany different types of networks aretechnically feasible. An analysis wasmade of different possible configura-tions of networks, and their possiblecosts were estimated. The book alsooutlines an organizational structurewhich would be able to operate sucha network. The last chapter is an in-formal proposal to a number of gov-ernment agencies who might contrib-ute to the funding of such a proto-type.

The basic concept of EDUNET isto encompass a number of activitiesalready underway in a limited fash-ion. Some of these are limited geo-graphically, regional networks. EDU-NET is conceived as a national orconceivably, ultimately, an interna-tional network. Some networks arelimited to a single medium, e. g., anetwork for programmed instruction,a television network, or a network forteletype transmission of bibliographi-cal information. EDUNET is devisedon the assumption that all mediamust be employed, each when mostappropriate educationally. From thetechnical point of view, it is possibleto operate multiple media on thesame bandwidth, switching from oneto another as required. EDUNET isalso intended to' involve all disciplineswithin the university. Some of EDU-COM's projects are for one field, suchas our plans for networks in bioagri-culture and biomedicine. Such en-

32

3 1

deavors are useful ways to begin, butlimitation to a few disciplines is edu-cationally unsound. Other fieldswould feel neglected if only physi-cists, chemists, engineers, and physi-cians had access to the new forms ofinformation processing. EDUNET isalso devoted to achieving as soon aspossible on-demand access by the userto information, on-demand access, notonly to documents but to televisionprograms, computerized programmedinstruction, and all other media.

We propose to join a number ofcolleges and universities in an experi-ment lasting three years or morebut for a definitely limited period.This experiment would be designedto ineicate what is feasible and whatis not, what is worth doing and whatis not, what is too costly and what isreasonable in terms of dollars and hu-man effort. The network would beoperated with one staff and evaluatedwith another, to be sure that the op-erators do not bias the evaluation ofthe benefits the system provides tothe users.

We propose to establish an EDU-NET laboratory with a single directorsupervising a staff at three or four lo-cations. One location would be at theResearch Triangle in North Carolina,one in Ann Arbor, Michigan, and oneon one of the southern campuses ofthe University of California. A fourthlocation might be at an Ivy Leagueuniversity. The EDUNET branchlaboratories would be interconnectedwith four switching stations, possiblyat San Francisco, Denver, Chicago,and Boston. The entire interconnec-tion would constitute the kernel net-work.

The first year we would spend ap-proximately a million dollars recruit-ing a staff and deciding on such ques-tions as the organizational structure

for EDUNET, its operating policies,issues of compatibility for hardwareand software, the selection of hard-ware for the branch laboratorief,what materials would be put on thenetwork, the selection of long-lineservices, and how costs would be ac-counted for and charged to users. Inthe second year, budgeted at approxi-mately three million dollars, narrow-band interconnection of the points onthe kernel network would be estab-lished and one after another servicewould be put on it. By "narrow-band"we mean a bandwidth equal to ap-proximately 12 parallel telephone

channels.Whenever any service became op-

erative, any EDUCOM universitythat wanted to use it could do so. Itwould have to pay the long-linecharges to the nearest point on thekernel network. The university coulduse the service any way it wished to

on its own campus. If it had a remoteterminal system or computer system

which was incompatible with EDU-NET, it would have to make thetransformations required eitherthrough a black box or by program-ming. Conversely, if they wanted tocontribute a service prepared on theircampus to other universities throughEDUNET, they would have to putthe information into EDUCOMpati-ble format so that it could be carried

on the network.In the third year, we would expand

the bandwidth to permit two-waycolor television. The cost would beabout six million dollars. We may beunderestimating costs. That dependson how complex our services become,

unanticipated operational problems,and inflation. We are attempting,however, to keep the costs as low aswe possibly can so that there is a rea-sonable chance to raise the necessary

33

funds. The universities will contributetheir appropriate share of the fundsrequired. The government will not beasked to provide the total amount.The universities will pay for theirown terminals, for the long-linecharges to interconnect with the ker-nel network, and for some services.

Many of the services we put on thenetwork will be prepared by govern-mental or nonprofit private organiza-tions. We have made tentative ar-

rangements with quite a number ofsuch organizations to use their serv-ices. Such inputs will of course bemuch less expensive to us than if wehad to prepare them. A considerableamount of machine-readable informa-tion is available now to go on such anetwork if arrangements can bemade.

The Office of Science and Tech-nolo ;y has taken cognizance of ourprop isal, has assigned responsibilityfor t to the Department of Health,Education, and Welfare, and underHEW to Dr. Louis Bright, AssociateCommissioner for Research in the Of-fice of Education. On various occa-sions he has called together interestedagencies to discuss the EDUNETproposal.

It seems probable that EDUCOMmay be requested to extend the plans

for another less ambitious network.We have proposed EIN, the Educa-tional Information Network, to in-clude certain aspects of EDUNET.

We have had a discussion aboutEDUNET with a member of theWhite House staff. On November 7,1987, the President announced hissupport for the development of "net-works for knowledge". In the Presi-dent's message on education to the1968 Congress, he recommended theNetworks for Knowledge Act of 1968,which is Title 9,of the Education Bill

of 1968, H. R. 15067. This proposesnetworks almost identical in principleto that outlined in the book "EDU-NET." Hearings on this bill were heldseveral weeks ago by the SpecialSub-committee on Education of theHouse of Representatives and the billis now being reviewed in the Senate.

I have mentioned EIN, the Educa-tional Information Network. EDU-COM is striving for the ultimate de-velopment of an embracing programlike that of EDUNET. At the sametime it is important, however, to carryon networking activities relevant tothe missions of various funding agen-cies. We have a contract with the Na-tional Library of Medicine to plan aNational Biomedical CommunicationsCenter which may involve network-ing in that field. We have a contractwith the National Library of Agricul-ture to develop a plan, in associationwith the EDUCOM landgrant univer-sities, for information transmission inbioagriculture. We have an agree-ment with the Dental Health Centerof the United States, when funding isavailable, to work in the area of adental health information system. Wehave a grant under consideration bythe Office of Education jointly withthe National Science Foundation forthe development of EIN, a networklimited to one of the media, i.e., com-puters. Work on the Educational In-formation Network is expected soonto begin according to the followingplanned chronology:

Phase I1. Engage a limited group (up to

20) universities in agreement permit-ting remote accessing of their com-puter facilities. ( More than 20 univer-sities have already volunteered tocooperate.)

2. Establish standard documenta-tion format for a description of com-

3 3

-*.. SCCM*AT

34

putational facilithproduce a director:tional facilities for

3. Establish tecfives at each particact as liaison andfion about operatio

4. Set up a reato identify usage pand further plannir

5. Establish an (supplement the intory.

Pha,s6. Establish a mi

mission amongwith their indivistraints.

7. Develop a syreferencing betweeputm at participa

8. Generate an(bases for expanded

We believe thaare being spent iCOM universitiesgrams for local coalready on comput,sities. This is trumany sorts fostruction, for datcomputational usamong others. Tocosts of such dupenable a user on (into a computer cwhich already hassires, so permittirfrom the work awhere.

I have been ablsome of our presthose only brieflythere will be apjprograms to the htient-record systeneducation, researcivital that all scient

al people be able to gain rapid accessto relevant information and that ourinstitutions be able to share resources.Networks can help in this. They canenable our universities to become asociety of universities, to establish inour nation universities without wallsthat penneate many activities of ourcountry. They can give us democracyof access to information. Not only inBoston, New York, Chicago, and SanFrancisco will users be able to obtainspecialized information in depth, as isnow true, but a new college or uni-versity wherever it is, a small institu-tion in the Rocky Mountains or a fi-nancially weak Negro college in Ala-bama can have as good or better ac-cess to information as Harvard does.Networks can aid in improving thegeneral quality of instruction, re-search, and professional serv icesthroughout the nation.

Discussion PeriodsQuestion: Has EDUCOM devel-

oped relationships with industry?

Dr. Miller: We have received no-strings-attached grants for EDUCOMactivities from the General LearningCorporation from IBM, and in smal-ler amounts, from other corporations.We also have received a grant fromRCA to support a conference. Wehave six members-at-large on theBoard of EDUCOM and three ofthose members are from industry.Each of them is an individual who isdistinguished in education, for exam-ple, Francis Kleppel, who is wellhown as an educator and is alsoChairman and President of GeneralLearning Corporation; and Carrol I.Newsom, Vice President for Educa-tion of RCA, former President of NewYork University. We have also atpresent a representative from IBMand one from General Electric. Atvarious EDUCOM conferences repre-

35

34

sentatives of the publishing, long-line,hardware, and broadcasting indus-

tries, have participated. While EDU-COM is planning a prototype experi-mental network, it is not at all surethat in the long run it is the responsi-bility of the universities to run suchnetworks. After we are clearer hownetwork services can help universi-ties, some industry or industries maywish to provide the services on a rea-sonable profit-making basis. Thatwould be quite acceptable to EDU-

COM.Question: Has EDUCOM devoted

attention to problems of compatibilityamong local and regional networks?

Dr. Miller: We may expand theEIN study to determine the technicalcharacteristics of all the regional edu-cational networks, for computers, li-braries, television, and so on, in an ef-

fort to ascertain to what extent they

are compatible. Millions of dollars oflocal, state, and federal funds are go-ing into regional networks of varioussorts. So far as I know, no one region-al network planning group is takinginformed notice of the technical char-acteristics of another in order to becompatible. EDUCOM may attemptfirst of all to learn the characteristicsof present networks. Then, when wehave developed, in the light of thesefacts, suggestions about what mightbe appropriate standards, we will

propose them through proper techni-cal channels. We are conscious of thefact that if you establish compatibili-ty regulations too soon you may re-strict proper development. On theother hand, sometimes decisions mustbe made or chaotic waste occurs. Ipersonally believe we are near thetime when we must make such deci-sions. We have talked to the Bureauof Standards about such matters, as

well as to other agencies involved in

setting standards. While some stand-ards for television, for exampleare now well established, in otherareas there are none at all. We wouldobviously follow Library of Congressor Library of Medicine standards orothers set by responsible bodieswhenever they exist. In areas wherestandards have not been formulated,some group must make the decision.Large hardware and software invest-ments are being made by universitiesin the current year. If these products

obtained are so incompatible thatWisconsin cannot take advantage ofdevelopments at Cornell or ClevelandState cannot benefit from work atCalifornia, there will be a great lossto all of us. There is danger of a de-bacle like the color television hasslebetween CBS and NBC some yearsago or the battle among the makers ofdifferent sizes of long-play records ortape cartridges. A solution must befound soon and EDUCOM is awareof the problem.

COMPUTERS IN UNDERGRADUATEMEDICAL EDUCATION

HILLIARD JASON, M.D., ED.D.

Undergraduate medical education inthis country is in a state of re-examina-tion and revision in proportions withoutprecedent. "Efficiency" and "effective-ness" have become watchwords of cur-riculum committees. Patterns and tech-niques of instruction which had beenhallowed by decades of use are beingmodified and even jettisoned, and thesearch is on for new instructional modal-ities which hold the promise of gettingmore done in less time.

The growing sense of need for bettereducational methods has been paralleledby an increasing availability of potentialsolutions. Within the past decade wehave come to understand the process ofinstruction more precisely and techno-logical aids to the process have ap-peared in increasing profusion. Pro-grammed instruction, teaching ma-chines, automatic cartridge movie pro-jectors, simplified audio recorders andtelevision are all part of the growingtechnology of education. Because of itsfar greater flexibility, potency and cost,the computer represents the culminationof this trend.

To understand the promise and theproblems computers might bring tomedical education, we must understandsomething of: (1) the process of intro-ducing a major innovation into an edu-cational system, (2) the current educa-tional requirements in medicine, and(3) the particular characteristics of the

computer which may or may not fit intothese educational requirements.

Innovation in EducationAn innovative device of the scope of

the computer can be a mixed blessing.We have recently begun to enter a newera in medical education: one in whicha half century of educational conserva-tism is giving way to a receptivity for,indeed a hunger for, new and promisingapproaches. As much as many of us re-joice at this venturesome spirit, we mustbe wary of its potential pitfalls. Mostworrisome of all is the capacity of newdevices to be mistaken for new ideas.The act of accommodating old instruc-tion to new packages, such as televisionor the computer, can prove to be worsethan mere maintenance of the statusquo. The glossy new packaging tends todeceive both the student and the teach-er into believing that something new,and therefore better, is happening. Inactual fact, what often happens is thatconsiderable energy and money are in-vested in creating a situation that iseven harder to change than was theconventional instruction it is mimicking.

We have repeatedly seen, for exam-ple, the poorly delivered stultifying lec-ture become even more so on the TVmonitor. With sadness I must report thatthe equivalent has begun to occur withthe computer, in some people's hands.One can now find examples of electronic

°Director, Office of Medical Education Research and Development, College of Human Medicine,

Michigan State University, East Lansing, Michigan

37

43

reincarnations of the pompous, arrogantprofessor whose instructional strategy isthe demand of compliance rather thanthe promotion of independent skills inthe student. The computer generatedreprimand of "Nonsense! You shouldknow better!" fails to help the studentlearn what he did wrong or where to gonext. Simply put: bad instruction isn'tmade good by inclusion in a new medi-um. The essence of these miscarriages ofmodern methods is that they are at-tempts to provide solutions, withoutawareness of the nature of the existingproblems: they are accommodations ofconventional instruction to availabletechnology, when continuing modifica-tion of the instruction itself and adapta-tions of the technology are needed. Foreducational innovation to be meaning-ful, it must derive from a sound under-standing of educational requirements, sothat available technology can be proper-ly modified and exploited, and therebyprovide worthy solutions.

Requirements of Medical EducationA. What We LearnI will turn now to some requirements

of an effective medical education: first,in terms of what it is that students mustlearn.

The most frequent and most desper-ate plea one hears today in medicine isthat we must find ways for dealing withthe inundation with which we are al-legedly faced, as a consequence of the"information explosion." Statistics are re-peatedly cited on the multiplying num-ber of journals and journal articles as ev-idence of this trend. There are two sidesto this problem which deserve our atten-tion. They are neatly highlighted by thefollowing quotation:

"Those who have occasion to enterinto the depths of what is oddly, if gen-erously, called the literature of a scien-tific subject, alone know the difficulty of

38

emerging with an imsoured disposition.The multitudinous facts presented byeach corner of Nature form, in largepart, the scientific man's burden today,and restrict him more and more, to anarrower and narrower specialism. Butthat is not the whole of his burden.Much that he is forced to read consistsof records of defective experiments, con-fused statements of results, wearisomedescriptions of detail, and unnecessarilyprotracted discussions of unnecessaryhypotheses. The publication of suchmatter is a serious injury to the man ofscience."

These timely remarks are taken fromJohn Langley's presidential address tothe Physiology Section of the Royal So-ciety, in 1899 (1). It seems that the in-formation explosion, and its accompany-ing "garbage explosion," began sometime prior to the present decade.

Most of us would willingly concedethat much of the seeming explosion ofknowledge is in truth, a dihition withtrivia and irrelevancies. There has been,however, an undeniable and very con-siderable expansion of our understand-ing of life processes. I challenge, howev-er, the usual implications that are de-rived from this expansion. Rather thancomplicating the process of learning, Icontend that the growth of knowledgehas actually simplified the process. Iwill turn to some other quotes to helpmake the point.

"I know that I am shockingly igno-rant. I could take exams in college butcould not pass any of them. Worse thanthat: I treasure my ignorance; I feelsnug in it. It does not cloud my nsivete,my simplicity of mind."

This is not the statement of a SanFrancisco "flower child," it is the obser-vation of one of our most distinguishedscientists, the Nobel laureate, AlbertSzent-Gyorgyi (2). It was said in a con-text that should become clear from the

Ii,

following additional quotes:"It is widely spread opinion that me-

morizing will not hurt, that knowledgedoes no harm. I am afraid it may. Deadknowledge dulls the spirit; fills thestomach without nourishing the body."

He points out that we have books andlibraries "to keep the knowledge inwhile we use our heads for somethingbetter." And it is our instructional chal-lenge to help students learn to use theirheads for "something better" than sim-ply storing information. A major point ofthis paper, to which I will shortly re-turn, is that the computer can help us dobetter jobs of both storing informationand teaching students to use their mindsfor "something better." First, let me lin-ger one moment longer with Szent-Cyorgyi, to emphasize a notion that willhelp clarify the point.

"Science tends to generalize, and gen-eralization means simplification. Myown science, biology, is today not onlyvery much richer than it was in my stu-dent days, but is simpler, too. Then itwas horribly complex, being fragmentedinto a great number of isolated princi-ples. Today these are all fused into onesingle complex, with the atomic modelin its center. Cosmology, quantum me-chanics, DNA and genetics, are all, moreor less, parts of one and the same story

a most wondrous simplification. Andgeneralizations are also more satisfyingto the mind than details. We, in ourteaching, should place more emphasison generalizations than on details."

Put in other words, the expansion ofknowledge has left us with a lighter,rather than a heavier, burden. We arefar better equipped than ever before tohelp our students come rapidly to gripswith the central, organizing issues of ourdisciplines. We can provide the concep-tual base which permits the skills of themanipulation of biowledge, the skills ofacquiring, validating and using knowl-

39

38

edge, rather than the mere possession of

knowledge, io become the mark of theeducated man in medicine. It is the actsof evaluating and generating knowledge,more than retaining it, that seem propergoals for our instructional programs.And it is in this transformation of goals,

not in the adoption of new devices, thatthe heart of educational innovation lies.

B. How We LearnLet us turn from what students should

learn to how they learn. Among themany conditions of effective learning,two requirements are paramount: First,the student must practice, during learn-ing, the very things he is to be able todo as a result of learning. That is, if heis to be able to distinguish abnormalfrom normal heart sounds, he must prac-tice that vary act. If he is to be able tosolve clinical problems, he must practicethe various steps, the components, ofproblem solving. Second, the studentmust have frequent and reliable infor-mation on the appropriateness of theprogress he is making: he must have"feedback." To illustrate: if he is to

learn to swing a golf club, he must notonly get to see where the ball goes as aresult of his efforts; he will make hismost efficient progress if he can see amovie or videotape of his actual swingand he is helped to see the crucial com-ponents of the total skill, such as the po-sition of his head, the line of his shoul-ders, his stance, and so on. Similarly, ifthe student is to learn to solve medicalproblems, he needs to know the correct-ness of his solutions, hut he also needsto know his competence at managingthe crucial components of the process;such as problem sensing, problem for-mulation, information search and resolu-

tion. These skills have been identifiedand elaborated by Shulman (3).

While the earning of intellectual skills,such as problem solving, is far morecomplex than the learning of mechanical

skills such as golfing, we have been con-siderably less effective than golf instruc-tors in providing the exquisitely impor-tant ingredient of meaningful feedback.

To recapitulate, I suggest that amongthe central challenges of current medicaleducation are: (1) developing tech-niques for helping students learn to ef-fectively manipulate knowledge, (2) en-abling them repeatedly to practice thesemanipulations, and (3) providing fre-quent and reliable feedback on theirsuccess at the overall process, and at itscomponents. Computers have a uniqueset of characteristics which, if properlyexploited, can fulfill these requirementswith remarkable effectiveness.

Characteristics of Computers

The specific characteristics of moderncomputers and the ways in which theyfunction, have been well described in arecent publication by Caffrey (4 ), in-tended primarily for technically unin-formed readers. The general characteris-tics with which I will here deal arethose matters which provide the com-puter with its special potential as well asits limitations, as an educational device.

Among the com ru t er 's greateststrengths are the SPUD with which itcan search its memory, undertake a cal-culation or solve a logical problem, thePATIENCE with which it will systemat-ically repeat questions or otherwise dealwith refractory students, the nearly un-limited capacity of its MEMORY forstoring all manner of information, theIMPARTIALITY with which it will dealwith different students and changingcharacteristics within the same student,the consistent LOGIC with which ithandles sequential information or intri-cate problems, and its ADAPTAMLITYto wide variations in the manner of pres-entation of the same problem throughthe availability of virtually unlimited ca-pacity for branching to alternate possi-

40

39

bilities. In the process of performing allthese functions, it can keep completerecords of the entire proceedings. As wewill see shortly, this is a singularly im-portant function for educational pur-poses.

Among the more serious of the com-puter's limitations in performing thekinds of educational challenges it tendsto be given are its FLEXIBILITYbeing limited as it is in its responses tothe finite repertoire which its programnecessarily provides, its essential lack ofORIGINALITY, in comparision to thehuman teacher with which it is so oftencompared, its FALLIBILITY in terms ofthe electromechanical breakdowns towhich current computers remain prone,and its essential lack of PERSONALI-TY, or patently synthetic programmedpersonality, which remains a somewhatjarring characteristic to many studentswho are otherwise caught up in the hu-manoid characteristics of the interaction.

The process of adapting the computerto the instructional requirements inmedicine must include the extraction ofmaximum advantage from its strengthsand careful accommodation to its limita-tions.Computers in Undergraduate MedicalEducation

Before discussing the computer as aninstrument of direct instruction, I willbriefly mention some of the ways it canincrease the efficiency and capabilitiesof administrative functions which sup-port the overall educational system. Al-though they are not seen as part of thepurposes of the present paper, they arementioned for completeness. The proc-esses of student record keeping, classand patient scheduling, educationalbudget management and the develop-ment of a reference memory on a varietyof relevant subjects are all parts of thecomputer's potential contribution to ed-ucation. They have been discussed by

Caffrey (4). The exceedingly importantfunctions of library maintenance andmanagement have been described byBrandt (5) and will be elaborated uponin this conference by Dr. Martin Cum-mings.

The remainder of this paper will bedevoted to those educational applica-tions in which the student receives in-struction through direct contact with thecomputer, using a remote terminal. Thisconsists of a typewriter-like apparatuswhich permits both the student and thecomputer to ask and answer questions ina fairly free exchange. There may be, inaddition, a television-like tube for thedisplay of printed information or dia-grams, as well as computer controlledslide projectors, movie projectors, andaudio recorders, for providing the stu-dent with sensory inputs of varioustypes. I will briefly examine the fourmajor types of educational uses withinthis setting.

A. Information and ConceptsConventional medical education is so

information oriented that it is no sur-prise to find that a fair proportion ofearly computer based instructional pro-grams have been primarily informationoriented. A fairly typical illustration ofthe strategy employed is represented bythe model adapted from Spolsky (6) inFigure 1. This mudel, it will be noticed,takes advantage of the adaptability, log-ic and patience of the computer to sys-tematically check, and if necessary cor-rect, the progress of the student in ac-quiring a body of information. It shouldbe clear that the computer is fully capa-ble of drilling students in the acquisitionof information. Information acquired inthis way, however, will generally notlast long. Information of the complexityrequired in medicine, paradoxically, isacquired best when it .is not the primarycenter of attention, but is part of appli-cations that are meaningful to the stu-

41

4it 0

dent. As discussed earlier in the paper,information that is not alive and vividfor the student is best left to books orsome other repository.B. Intellectual Skills

Essential in its relevance to the gradu-ate physician, and hence to the studentphysician, are the skills of identifyingand solving problems, generating andtesting hypotheses, and formulatingjudgments. The characteristics of thecomputer make it possible to present thestudent with repetitive instances ofthese procedures for practice and mas-tery. It is here, it seems to me, that thecomputer holds its greatest potential inundergraduate medical education. Re-turn to Figure 1 and imagine, if youwill, a problem oriented, rather than aninformation oriented, program. The er-rors that the students here make areprocess rather than content errors, andthe remedial assistance they are givendeals with the components of the intel-lectual skills being learned, rather thanwith facts being acquired. For example,they might be alerted when they drawan unwarranted inference or arrive at ajudgment with insufficient evidence.And, most important of all, the comput-er's capacity to retain a total record ofall transactions in which it engages, en-ables the student to be provided withthe cognitive equivalent of a video-tapeof his thought processes. In fact, thecomputer print-out of the problem-solv-ing process in which the student has en-gaged is far more flexible than avideo-tape. It is more like a map whichcan be scanned in its entirety and com-pared to other maps, for critical evalua-tion. Much like the golfer who makesstartling progress through being con-fronted with his own swing, the medicalstudent learning intellectual skills, whenprovided with the feedback of theprint-out, can make unprecedentedstrides in areas previously barely

t.,t

IFIR

ST

DIA

GN

OS

TIC

FR

AM

E

1

1

PR

ES

EN

T R

EM

ED

IAL

RO

UT

INE

IR

EIN

FO

RC

EM

EN

T

I

1F

IGU

RE

1

GO

SE

E IN

ST

RU

CT

OR

I

touched by conventional medical educa-tion.

C. Technical SkillsThe flexibility of the computer, com-

bined with a multimedia terminal, per-mits the student to be presented withwide varieties of EKG tracings to be in-terpreted, slides to be read, or heartsounds to be deciphered, to mentiononly a few of the possibilities of instruc-tion in technical skills where student re-sponses can be immediately judged andrepeated practice provided. An evenmore sophisticated possibility is thecreation of complex simulation situa-tions, such as "Sim I," developed by Drs.Stephen Abrahamson and J. D. Densonfor the training of anesthesiologists.

D. EvaluationThe capacity of computers to keep a

complete record on all its transactionswith students permits an unusually com-prehensive and elaborate evaluation,both of student performance and in-structional effectiveness. The singularlyattractive possibility which emerges isthat of computer-monitored instruction.In a way we have never been capable ofdoing in the past, student progress canbe assessed from day to day, rather thanterm to term, and instructional adjust-ments made, literally, as needed. Theextent to which this can elevate our effi-ciency can, at present, barely be imag-ined. In addition, the memory and adap-tability of the computer permits the de-velopment of a store of great ranges ofevaluation materials, which students canuse at any time to assess their own prog-ress along many dimensions of any disci-pline. Still further, formal accreditation-type evaluations can be undertakenmore flexibly and more reliably with thecomputer. An imaginative new comput-er-based technique for evaluating stu-dents of large differences in ability, in amanner that exposes them to a uniqueset of questions which are maximally

43

adapted to their level of performance,has been described by Chauncey (7).

CONCLUSIONSThe computer, without question, pro-

vides the potential for an enormoustechnological leap forward in education.Its long range promise and conse-quences can barely be glimpsed, if rec-ognized at all. Many of the steps, how-ever, which must now be taken to helpus profit most from this potential, canreadily be seen.

Most important is the need for us toidentify, as explicitly as possible, thereal purposes of medical school instruc-tion. We must then define, as preciselyas we are able, the characteristics andcomponents of the competencies ourstudents are to acquire. These must beformulated in terms that duplicate orparallel the circumstances in which theywill actually be applied following grad-uation, so the practice we provide ourstudents in school will be maximally rel-evant and real. Provision must also bemade for effective feedback to the stu-dents on their success in acquiring thesecompentencies.

A burning question in the minds ofmany is: if we succeed in properly ex-ploiting the potential of the computer,will the live teacher be displaced? In tanobservation that is equally relevant tocomputers, B. F. Skinner has said that,"Any teacher who can be replaced by ateaching machine, deserves to be." Thepoint is this: the computer can, indeed,do certain things better than the liveteacher. It was more than a half centuryago that Thorndike observed, "a humanbeing should not be wasted in doingwhat forty sheets of paper or two phon-ograph records can do. Just because per-sonal teaching is precious, and can dowhat apparatus cannot, it should besaved for its peculiar work." (8) Thechallenge before us is not how to sup-press computers so as to permit teachers

to go on doing those things they nolonger need to do. The challenge is,rather, to exploit computers to theirmaximum and, at the same time, work toredefine the human teacher's role so thatwe may take full advantage of his po-tential, which no machine of the foresee-able future will possibly duplicate.

No innovation of this potential magni-tude is without risk. Shall we wait untilall the data are in and move forward sodelicately and with such caution that allfalse starts and errors will be avoided, orshould we acknowledge the enormouspromise now available to us and forth-rightly and vigorously take some

REFERENCES1. Fletcher, W. M., J. Physiol. (London),

61:1-27, 1926,2. Szent-Gyorgi, Albert, "Teaching and the

Expanding Knowledge," Science, 146:1278-9, Dec. 4, 1964.

3. Shulman, Lee S., "The Study of IndividualInquiry Behavior," Presented at the Amer.Psychol. Assn., Washington, D. C., Sep-tember 2, 1967.

4. Caffrey, J. and Mosmann, C. J., Computerson Campus, Washington, American Councilon Education, 1967.

5. Brandt, E. N., Jr., "Electronic Computersin Medical Education of the Future," inClinical Anesthesia, 1:125-36, 1966.

0. Spolsky, B., "Some Problems of ComputerBased Instruction," Behavioral Science,11:487-96, 1966.

7. Chauncy, Henry, "Some New Approachesin Testing," J. of Med. Ed., 41:NO. 7, Part2:25-32, July, 1966.

8. Thomdike, E. L., Education, A First Book,New York: The MacMillan Company, 1912,p. 167.

ADDITIONAL BIBLIOGRAPHY1. Bowden, D. H., Computer Aided Instruo-

tion in Pathology. Canadian Medical As-sociation Journal, 97:739-42, 1967.

2. Brown, M. C. Electronics and MedicalEducation. Journal of Medical Education,38:270-81, 1963.

44

4 3,,

chances? Let me answer the question byanalogy. Everyone seems to rememberthat Babe Ruth held the all-time homerun record for many years. Very fewpeople seem to remem'ber that for al-most as many years he held the all-timestrikeout record. It seems to me that inall areas of human endeavor, theachievement of greatness comes only

from taking calculated risks. Let us takesome risks, because the promise of com-puters seems without precedent in thehistory of education, and, intelligentlyused, should provide us with the instruc-tional equivalent of a series-winninggrand-slam home run,

3. Dryer, B. V. Thinking Men and ThinkingMachines in Medicine, Journal of MedicalEducation, 38:82-92, 1963.

4. Entwisle, G. et al. The Use of the DigitalComputer as a Teaching Machine. Journ41of Medical Education, 48:803-12, 1963.

5. Feurzeig, W., et al. Computer-AidedTeaching in Medical Diagnosis. Journal ofMedical Education, 39:746.54, 1964.

6. Fonkalsrud, E. W., et al. Computer As-sisted Instruction in Undergraduate Surgi-cal Education. Surgery, 62:141-7, 1967.

7. Harless, W. G. The Development of aComputer Assisted Instruction Program ina Medical Center Environment. Journal of

Medical Education, 42:139.45, 1967.8. Kirsch, A. D. A Medical Training Game

Using a Computer as a Teaching Aid.Methods of Information in Medicine, 2:138-143, 1983.

9. Long, J. M. et al. Experience TeachingComputer Programming to Medical Stu-dents. Journal of Medical Eduanion, 40:65740, 1965.

10. Overstone, J. A. Computer Asststed In-struction in 'Undergraduate and Post Grad-uate Medicine. Medical Journal of Aus-tralia, 2:487-91, 1968.

11. Swets, J. A. a al. Computer Aided Teach-ing. Science, 150:57241, 1965.

12. Suppes, P. The Uses of Computers in Edu-cation. Scientific American, 215:207-220,September, 1966.

SUMMARYCOMPUTERS IN UNDERGRADUATE MEDICAL EDUCATION°

Panel: Hilliard Jason, M.D., Ed.D., Chairman, Stephen Abrahamson, Ph.D., Alan Kaplan, M.D.,Theodore Peterson, Ph.D., Lee S. Shulman, Ph.D.

It Ls not possible to give credit in asummary like this to everyone's favoriteidea. As a matter of fact, it is difficult toinclude the most salient points of athree-hour discussion. However, this pa-per will attempt to give at least an over-view of the major ideas discussed.

Our approach was first to take a looket the present state of the use of com-puters in undergraduate medical educa-tion. Dr. Theodore Peterson reported asto the work underway at Harvard Uni-versity, same generalities based on theirwork and their summary of the work ofothers. Dr. Lee Shulman helped us lookinto the future to some new ways com-puters might be used in achieving someobjectives, particularly in the area of de-veloping the processes of medical in-quiry. Dr. Stephen Abrahamson helpedus stay reality-oriented with his very aptand effective reminders of the ap-proaches that are necessary to evaluateproperly what we are doing as we do it,and Dr. Alan Kaplan of the Bureau ofHealth Manpower provided perspectiveas to the hopes, wishes and plans of thefederal government in the area of sup-port of computers in medical education.In order to bring some coherence to arather free-flowing and not always sys-tematic discussion, eight general conc u-sions derived from the discussions willbe presented. In addition, Dr. Shulman'scomments will be presented as an ap-pendix to this summary.

The first conclusion which can bestated with some acronymity, is thatcomputers in undergraduate medical ed-ucation are not yet in what might becalled an implementation phase. Anyschool that does not now have a com-puter should not feel that they are some-how missing out on an importantcontribution to their education program.In fact, this leads to the second generalconclusion, which is, that the primaryjustification for having a computerwould be for a few innovative and evenunique studies of its utility and effec-tiveness in medical education.

The third conclusion which ties intothe first two, is that the issue of cost inthe use of computers is a complex but aseemingly important one, and thereseemed to emerge two different posi-tions regarding cost in the ways comput-ers might be used in undergraduatemedical education. When computers areused to do something that is alreadybeing done, then cost was an exceeding-ly important consideration. The ratio ofcost to effectiveness was one of the cen-tral issues in determining whether whatwe now do can possibly be done differ-ently or better with a computer. Butthere was a second position and that isof the possibilities that computers bringthat may enable us to do things wearen't doing at all now and may enableus to realize instructional objectives thatare not available now. In areas like this,

°Presented by Hilliard Jason, M.D., FAD., Chairman, at the closing plenary session of the con-

ference.

45

cost becomes a far less important con-sideration. That is, if we can use thecomputer and our studies of it to help usproduce a physician of higher qualitythan we have produced before, then costis no longer an issue. Our dedication hasbeen that if we can produce somethingreally worth producing, cost becomes asecondary issue.

A fourth general conclusion is that thetechnology of the computer seems to beconsiderably ahead of our readiness totake full advantage of it. The most cen-tral and most pressing need that was re-peatedly identified was the urgencywith which we need a far better delinea-tion of our objectives in undergraduatemedical education. These are not thebroad, philosophic, abstract program ob-

jectives of trying to produce good physi-cians; these are the hard-to-define butexquisitely important instructional ob-

jectives that commmunicate what are in-deed the specific competencies that agraduate will have, if we are to regardhim as a worthy graduate of our school.These have never been defined, and ashas been said before, if you're not surewhere you're going you might end upsomewhere else. The need for precisionin defining objectives does not requiregovernment support, it does not requirehardware; it does require the conscien-tious dedication of medical educators tothe task of working together and comingto grips with what it is a physician isgoing to be able to do. Not only what heis going to be able to know, but what heis going to be able to do, if we have con-fidence in him and in the quality of ourinstruction. The important point is, thatwhether we go about doing that expli-citly or not, we have done it. Withoutdefining objectives explicitly, we havedefined all sorts of objectives implicitly,just by doing what we do when we as-sign a particular class to a particulartopic. That is a definition of objective.

48

MOM 1.41aMillar. erMl

By failing to be explicit in objectives,run a terrible risk of which we wereminded several times on the panel. Tlis, the risk of achieving some objectiat the expense of more important onthat by helping students work in a pticular area we might be failing to hthem learn how to work in another arFor example, if we have students divery significant portion of their inpendent learning on a computer,we perhaps running the risk of failinghelp them with the important skillslearning how to learn without a comier? There are more important questithan that, and these relate to our crent examination system. We commtcate to students most effectively wour objectives are by the exams wesign. Another way of communicatingfectively our objectives is by investingreat deal of effort into what we putside a computer. If we convey a lol

ephemeral details and very short-liinformation because that's what isiest to put into a computer or easiesexamine on paper and pencil tests, twe are communicating to them furmentally that learning is not a very rvant activity, and that the act of parpating in medical education is a g;which you play in order to get a degThen you end up with one of the gtragedies of medical education in Auca; the production of generationsphysicians who are committed to thetion that learning is something theyspend the rest of their career avoidIf we don't design our programs, oul

tal educational programs, and our cputers in a way that helps physicdevelop the skills and dedication totinue to modify themselves the restheir lives, anything else we do is oimportance.

The fifth general conclusion thatin very closely with the last, is thaftuition just isn't enough in educati

planning or computer usage in medicaleducation. Intuition has led to a pro-gram of medical education in this coun-try that can safely be described as rigidin the extreme. Despite the evidence oftho remarkable diversity among medicalstudents, we have somehow intuitivelyfashioned the program that tries to pre-tend that everybody learns in the sameway, at the same speed, and is interest-ed in the same things. All of this flies inthe face of everything we know aboutpeople. So that if we are going to usecomputers intelligently, we need to putaside intuition because it has faitod us,and we have to start assembling data.We have to build into our computer us-age a very generous allocation of funds,time, and effort; to start a process ofevaluation of what we are doing; and todevelop a research base that will guideus in the future. A specific suggestionfinding favor was the idea of encourag-ing and developing regional efforts atresearch in medical education awl var-ious parts of medical education cate-gorical sub-parts of the total picture.

The sixth general conclusion would bethat it is inexcusable to undertake aproject today ir the use of computers inundergraduate medical education with-out a major investment in evaluation ofthe process.

The seventh major conclusion wasthat the promise of computers for un-dergraduate medical education is enor-mous, indeed. It brings more than justits own utility; it is having the heuristiceffect of forcing us to clarify our educa-

47

Fib

'I 0

tional objectives more precisely than woever have before, of promoting a re-thinking of our purposes and a re-think-ing of our methods that is long overdue.More significant still, besides being ca-pable of direct instruction and being ca-pable of direct evaluation of students,the computer is capable of storing itsown processes as they are conductedand becomes a remarkably effective toolin evaluating its own functions with ourdirection and with our design.

The eighth and final mafor conclusionis finding the proper balance betweenthe live teacher and the technology rep-resented by the computer, which is thegreatest challenge that we face. Ourchallenge is not only to use the comput-er intelligently, but to re-define imagi-natively the role of the live teacher be-cause so much of his time and effort inundergraduate medical education is nowdevoted to things we already know canbe done better by technology. If we aregoing to provide a quality medical edu-cation, if we are going to undertake theremarkable complex transformation of acollege student into a professional posi-tion, we have to provide him with effec-tive models of what it is we want him tobecome. What we need is the creationof live teachers who will embody, willnot just espouse but will embody thosethings we have decided we want ourstudents to become. This a machine cannever do and this is where the properrole of the live teacher begins to loe de-fined and this is where much work re-mains to be done.

INQUIRY, COMPUTERS AND MEDICALEDUCATION*

LEE S SHULMAN, PHI/

I speak to you as someone engaged inthe study of the manner in which peoplelearn, organize what has been learnedand employ that organized body ofknowledge in the service of solvingproblems. I label this process inquiry,after John Dewey. An examination of in-

quiry can provide educators concernedwith the teaching of future physicians asource of monitoring, evaluation andmodeling that will serve them exactly as

we have claimed the computer can servethe student. We say that it is importantfor the student to receive appropriatefeedback so that he has a constant per-ception of how closely he is approximat-ing some ideal model of how a physicianoperates. If we as educators are going topresume to program the computer toprovide that kind of service for the stu-dent, we need very clear ideas in ourown minds of what are the criterionmodels of medical practice, medical di-

agnosis, interviewing, history taking andphysical examination. These are themodels which the computer must ern-ialoy. We ought to specify these modelsbefore we desigm any program. I do notknow that we have accomplished this, as

yet.Types of Models

How do we generate this kind ofmodel? One approach, which we mightcall the experience model, is to makecareful observations of men and womenwho are identified by the community of

medical 'practitioners and scholars asbeing criterial physicians. These individ-uals represent in their behavior the fin-est kind of medical practice we canidentify. We then attempt to abstractfrom a careful analysis of what they dounder various conditions, and finallyemploy that abstraction as the model weput into our computer so that it, in turn,can now react and give feedback to themedical student who is learning. This isthe sort of thing that is done in the heu-ristic programming of a computer toplay chess or prove mathematical theo-rems. We do not have in the medical lit-

erature extensive studies of the actualbehavior of fine diagnosticians in suffi-cient detail to lead us to a model such asthis one. This kind of research will benecessary in order to make medical in-

struction significantly more effective.

The second approach is to developlogical models. We ask ourselves, nothow do good practitioners practice orgood diagnosticians diagnose, but ratherwhat appears logically to be the modelfor inquiry, be it medical inquiry or anyother laind of inquiry. The illustrativemodel I shall now describe, in simplifiedform, is derived from Dewey's philoso-phy of inquiry. It has directed my re-search efforts over the last six years.

A Model of InquiryWe begin with the stage of problem

sensing. Most real inquiry does not be-gin the way we represent it in schools.

°Presented as appendix to summary of panel discussion on Use of Computers in Undergraduate

Medical Education.

*Associate Professor of Educational Psychology and Medical Education, Cortege of Human Medi-

cine, Michigan State University.

48

4 7:

0.,

f.,

There, we present students with prob-lems to solve which have already beenvery carefully formulated. Rather, muchof inquiry in a natural setting beginswith the problem solver or inquirerbeing placed in a situation that is terri-bly ambiguous and undefined wherethe edges are blurred and the distinc-tions are not clearly made. His first taskis to sense where the problems lie and todistinguish what is problematic fromwhat is not. We all recognize that twoindividuals can encounter essentially thesame situation, which one of them willsee as problematic while the other willmerely nod and go on. Problem sensingis a very subtle kind of process. We dopainfully little to train people in prob-lem sensitivity. It may be, however,differences in the ability to sense prob-lems that distinguishes the medical thor-oughbreds from the horses that pull themilk wagons.

Once a problem has been sensed, itmust be put into words or categories.That is, it must be formulated. JohnDewey observed that a problem wellformulated is half solved. The manner inwhich we formulate any sensed problemdelimits the kind of resolution that weare capable of generating. A problemformulation determines the kinds of in-formation we are going to seek out inour search, the kinds of hypotheses wewill entertain or reject, and the kind ofsolution that we will admit as possible.Problem formulation is the second stageor aspect of inquiry.

The third is search. Search can beboth external and internal. We not onlyconduct search by asking questions ofpatients, performing a physical examina-tion, calling for certain kinds of 1Phora-tory tests and generally gathering infor-mation from the outside. There is alsointernal search. We have enormousstorehouses of information between ourears and, potentially, in our computers.

49

The question is how much of this infor-mation do we seek, and which of it dowe seek? The interaction of the internaland external information-gathering con-stitutes this total process of searching.The search is defined as the sequence ofoperations, questions, movements, frus-trations, revisions of tactics, and the like,that the individual undertakes in orderto transform the problem-as-formulatedinto a personally-felt resolution.

Problem resolution occurs when wedecide that we are finished and theproblem which initially keyed off the in-quiry is solved. In the studies we haveconducted of inquiry in teachers, wefind enormous variability among indi-viduals in what they will accept as a res-olution of a problem. This is closely re-lated to their formulation, but it is alsorelated to individual differences in per-sonality, values, intelligence and train-ing. Unfortunately, differences attribut-able to training are usually accidental,because rarely do we have a clear modelin mind when we engage in instruction.

It is very important to note at thispoint that the model of inquiry de-scribed above does not require a rigidsequential ordering of steps wherein all

inquiries can be subdivided into foursimple stages representing each of thefour above-named processes. The natu-ral process of inquiry has these fourparts as somewhat independent components. Any particular inquiry will lookvery much like a computer programwith its many loops and digressions. Thesteps of inquiry act as the basic opera-tions in the program. Thus, it would notbe unlikely for a given subjeCt to sensethat a particular problem exists, proceedto formulate it in a particular way, initi-ate search activity in order to gatherdata concerning the problem, realize asa result of the initial search that hisproblem formulation is misbegotten, re-turn and reformulate the problem, while

doing so sense an additional problem ortwo, thus reformulate once again, and soon.Medical Inquiry

What happens when vou take thiskind of model and attempt to apply it inthe context of the medical situation?The diagram of Figure 1 suggests thatthe input of a specific case interactswith the organized body of knowledgepossessed by the physician, his long andshort-term expectancies of what he willencounter and the general strategies orheuristics he has learned, to generate asoutput the medical diagnosis, treatmentand evaluations. The full complexity ofmedical inquiry is surely not represent-ed in this diagram. The relations be-tween this model of medical inquiry andthe general inquiry mbdel are only sug-gested here. A far more detailed analysismust await much additional effort.

I think it is fair to say that the acqui-sition of organized bodies of knowledgehas been the major focus of discussionsin this conference, with the assumptionthat the computer is useful in makingmore efficient and effective the mannerin which these bodies of knowledge areacquired and organized. Not yet an-swered satisfactorily is: How do alterna-tive ways of organizing a body ofknowledge have different consequencesfor directing the process of inquiry?This is the name of the game in medi-cine, as it is in teaching and in othercomplex problem solving tasks. Weteach the organized bodies of knowl-edge, not to create little living librariesof our subject, but rather to make thesebodies of knowledge maximally retriev-able and useful in the service of inquiry.Is Socratic teaching, whether live or bycomputer, the most effeaive way of ac-complishing this? Is drill to the point ofoverlearning the most effective way?The most general question for the differ-ent aspects of the inquiry process we are

50

trying to teach in medicine is: Whichapproaches to learning are most likely toassist us in achieving our multiple goals?

The computer must be introducedinto the flow of medical instruction aspart of a systematic strategy of educa-tional development. First a map of ob-jectives must be drawn in very specificand operational terms. These will includeorganized bodies of knowledge, certainkinds of expectancies on the part of thephysician, and strategies of inquiry. Allthcse affect the conduct of medical in-quiry. The future impact of the comput-er on the practice of medicine is likelyto be reflected in a shift of instructionalemphasis away from the organized bod-ies of knowledge to the strategies ofknowledge-retrieval and inquiry. Next,the materials representing possible pathsto the terminal objectives must be devel-oped. Third, the materials and objec-tives must then dictate the combinationof software and hardware most ap-propriate to guiding our students intraversing the map. Fourth, the total in-structional system so designed must beimplemented under carefully monitoredconditions. Finally, the results of thatimplementation must be evaluated interms of the original map of objectives.This evaluation could lead to changedinstructional tactics, modifications in themap, or both. If we too quickly try tomove into the implementation stageswithout prior, or at least simultaneous,effort in the planning and design stages,we are likely to generate more chaosthan progress.Research in Medical Education

I shall conclude with an urgent plea Imake not only to medical educators butto educators of all kinds. For too longwe have allowed complex and expensiveeducational operations to be piloted bythe scats of our pants. We decide whatwe do on the basis of our reactions towhat was done to us. This can work in

Cit o-

EX

PE

CT

AN

CIE

S

OR

GA

NIZ

ED

BO

DIE

S O

FK

NO

WLE

DG

E

LEA

RN

ED

ST

RA

TE

GIE

S

INT

ER

AC

TIO

N O

FK

NO

WLE

DG

E. E

XP

EC

TA

NC

IES

& S

TR

AT

EG

IES

CU

ES

&

HIS

TO

RY

PA

TT

ER

NS

INP

UT

PR

OC

ES

SIN

G

1O

UT

PU

T

DIA

GN

OS

IS

TR

EA

TM

EN

T

PE

RC

EIV

ED

MO

NIT

OR

ING

&E

VA

LUA

TIO

NP

HY

SIC

AL

&LA

BO

RA

TO

RY

IXA

MIN

AT

ION

S

PR

OB

LEM

& P

RO

BLE

M&

SE

AR

CH

& P

RO

BLE

M

SE

NS

ING

V/ F

OR

MU

LAT

ION

IF (

INT

ER

NA

L. &

Tf R

ES

OLU

TIO

N

AN

DE

XT

ER

NA

L)

RE

FO

RM

ULA

TIO

N

FIG

UR

E 1

.A

Mod

el fo

r M

edic

al In

quiry

two ways. We can either repeat our ownexperiences or react against them anddo thc opposite. Both ate equally unten-able.

We must get beyond this level ofdoing things just because we intuitivelyfeel this is the best way to do it. Wemust attempt to study carefully theprocesses in which we are engaged. Iam not suggesting that people in medi-cal education rim to the studics of B. F.Skinner and try to extract his findingsand appl) them directly to medical edu-cation. That will not succeed, becausemedical education is not the same asteaching pigeons to play ping pong oreven 8-year-olds to count. Rather, weneed to create a discipline of the psy-chology of medical education that willbepr a relationship to the general psy-chology of education and learning arelationship that is parallel to the rela-tion medicine as a discipline bears to thebasic sciences. Clinical medicine is anintermediate discipline between the bas-ic sciences and direct medical applica-tion. Thc applied discipline of the psy-chology of medical cducation must haveanalogous characteristics. It does notnow exist except as a set of intuitively-derived dogmatic statements. How canthe creation of this field be accelerated?

The scope of systematic researchwhich is needed in medical education isgreat. In order to achieve that scope asquickly as possible, we must implementbroad programs of research in medicalteaching and learning, making best pos-sible use of the small number of current-ly active personnel in that field.

A strategy I would recommend foryour consideration is the following: Letus establish a series of Medical Educa-tion Research Laboratories around thecountry. They could be organized on ei-ther a regional basis, or through joining

52

together units which share interests butmay be in different parts of the country.Each laboratory would be dedicated toa specific research mission in medicaleducation and would draw upon the re-sources, facilities and research popula-tions of all its cooperating units uni-versities and hospitals. Such missionsmight include the specification and defi-nition of objectives for medical educa-tion, the development of evaluationmodels and evaluation instruments inmedical education, the uses of technolo-gy including the computer, the study ofthe medical inquiry process and studiesof the acquisition and use of organizedbodies of medical knowledge.

The Medical Education ResearchLaboratories must also assume a train-ing function in order to produce the re-search and development personnel need-ed to disseminate, implement and evalu-ate the research-based approacheswhich emerge from the laboratories.This medical education research effortcannot be left to individual medicalschools, but must be coordinated sys-tematically on a nationwide basis.

Ivan Pavlov is said to have told hisstudents, "Ideas and theories are like thewings of birds they allow man toclimb and soar to the heavens. But factsare like the atmosphere against whichthose wings must beat, and withoutwhich the soaring bird will surely plum-met back to earth." We are at a momentin the history of medical educationwhen we must buttress our myriadicideas and theories about the best formsof medical education with painstakinglycollected facts. In this effort the com-puter will be an indispensible tool.Without this research effort, the com-puter in medical education could be-come our newest enemy.

COMPUTERS IN CLINICAL MEDICAL EDUCATION

DONALD A. B. LINDBERG. M.D.°

There are a number of sourcesfrom which one learns clinical medi-cine. Computer systems have poten-tial roles to play in providing accessfor the student or physician to infor-mation. We will examine examples ofsuch systems.

The student or physician learnsabout medicine from four majorsources. These are: common senseand lay beliefs; by examining thefacts of patient care; by exposure topurely didactic information, and fromthe personal guidance and preceptsof his teachers.

Facts of Patient CareOne learns medicine from the op-

portunity to personally care for pa-tients or to see care given. This is avaluable, but limited experience. Onecannot see examples of each and ev-ery disease, treatment, outcome, andevaluation of such processes. Conse-quently one learns by drawing uponthe experience of a whole institution.A computer system can provide valu-able assistance merely by serving toretrieve patient records, to search thehospital files for records, and by per-mitting comparisons to be made 13e-tween different groups of patients ordifferent treatments or managementschemes. (1) (2)

In such cases the real learning isdone by the student reading andstudying the usual written hospital

record. The computer has merelyguided him to the correct record.

Inquiries into the records of labora-tory determinations can also be asource of learning. Systems have beenprovided so that the inquirer can seethe frequency distribution of resultsof a determination during past hospi-tal experience, and so place his pa-tient's results in a perspective of theinstitutional experience (3). The sys-tem provides the ability to retrieveany item of laboratory studies for anypatient, or alternatively to retrieveand display categories of results(such as, those patients with choles-terols greater than 300 mg%) or bycompounding categories (such as byalso requiring that the patients be ofa particular race, or sex, or that theresults have occurred during a certaintime period). It is also possible to re-quire in such an inquiry that the re-sults -of a second or of a thtrd deter-mination also meet certain specifica-tions. The net result of such com-pounded inquiries is merely the dis-play of data or the listing of patientidentification numbers no th in gcreative or innovative. Yet such a fa-dlity permits the student or physicianto ask questions of the mass of poten-tial "teaching data" which might oth-erwise be difficult to use via manualrecord room systems.

An additional benefit of the facility

*Director, Medical Center Computer Program, U niversity of Missouri Medical Center

4

53

to retrieve the results of laboratorydeterminations is provision to the lab-oratory director of added quality con-trol capability. He can, for example,specify very closely the format of re-ports, the requirements for release ofa report (i. c. tolerable limits" andspecial rules), the permissible preci-sion for reporting, the extent to whichquality control sera for a given batchrun may vary from the most preciseestimate, the extent to which aqueousstandards may vary from theirexpected curvilinear relationship. Thelaboratory director can take into ac-count the relationships between theresults of multiple simultaneous de-terminations in formatting his limits"systems (4). A simple example of thiskind of rule is the requirement thathemoglobin concentration and hema-tocrit bear a sensible relationship toone another when both estimates areavailable for the same blood sample.Computer systems for examination oflaboratory determinations can permitthe laboratorian to ask such questionsbefore erroneous reports creep intothe patient care system. When reportsdeviate from the limits" systems, theresident physician has an opportunityto re-examine the situation and ap-prove or disapprove the report. Suchconstant re-examination of premisesand performance as well as the crea-tion of new limits and rulm serves aneducational and training function forthe house officer and the laboratorydirector as %veil. At the very least,such systems direct ones attention tothe exceptional clinical case, ideallythe case requiring professional inter-vention and judgement.

Didactic InformationThe published medical and scien-

tifié literature is indexed by the Na-tional Library of Medicine, publish-ers of the Index Medicus. This cita-

tion data base itself can be searchedby the NLM computer system, andhopefully this system will be replicat-ed at many other medical centers(5).

There is now the potential to storeconsiderable quantities of informationon microfiche systems, which in turncan be indexed by computers (6). Atpresent one cannot (practicallyspeaking) transmit the images of mi-crofilmed text pages over long dis-tances. Perhaps image transmission(e. g. "Picture Phone") will in the fu-ture make a reality of the dream ofbrowsing through the library via thecomputer.

An experimental program at Mis-souri (called Consider) attempts toallow the student or physician to scanacross a broad spectrum of medicalknowledge. The data base employedis a magnetic tape version of theAMA Current Medical Terminology(CMT), kindly provided by Dr. Bur-gess Gordon. Our programs permitthe inquirer to type onto a computerterminal a statement cf the sizns.symptoms, or other findings from apatient. The system then searches theformatted statements of CMT anddisplays for the doctor a list of thenames of diseases in which the find-ings occur in the combinations speci-fied, and hence which must be con-sidered. Medical students have foundthe program challenging and ofteneducational. It is amazing how long adifferential diagnosis one can encoun-ter and how often the teacher has for-gotten much of what is contained inDr. Cordon's slim volume.

Recorded audio medical lectureshave been developed at the Universi-ty of Wisconsin (7). Under such asystem, the user can telephone to themessage center, request a particularlecture, and get the benefit of a care-

54

ful presentation without leaving hisoffice. The original audio tapes havebeen loaned by Wisconsin to variousRegional Medical Programs and othereducational groups. At Missouri theWisconsin tapes are utilized throughan automatic answering apparatus sothat the operation is totally unattend-ed. The critical ingredient in such asystem clearly is the necessary highquality of the lecture material. Com-puters enter the picture only on thetechnical side. Switching many in-

coming calls between hundreds (orpotentially thousands) of recordedlectures will be a computer job. Simi-larly, providing systems so that themedical user may learn of the exist-ence of an appropriate recorded lec-

ture may well be a future computerjob.

Research and AnalysisComputers have of course been uti-

lized for the statistical and/or mathe-matical analysis of research data formany years (8). Recently computersystems have made it possible to as-sist in learning clinical medicine byre-examining certain medical tech-niques. For example, use of Slack'sautomated patient history taking sys-tem (9) permits the doctor for thefirst time to repeat the process of ob-taining a patient history. The samecan be said of the interpretation ofelectrocardiograms by computer sys-tems (10). One probably cannot re-learn a fact or concept in quite thesame way (with or without a comput-er). Even so the advent of "ComputerAided Instruction" systems shouldloermit a far more careful study oflearning processes in those who learnthrough computer terminals than hasever been possible in the past.

55

It is beyond the scope of this paperto review the numerous excellent ven-

tures into true "computer diagnosis".I refer to the work of Lusted andLaney, Warner, Lodwick, Reichertz.and Moore and others. in all cases acomputer system is requiring certainobservations or measurements from

the physicians before attempting tocalculate a diagnostic likelihood. Theeffect on the user is to demand great-

er precision and consistency than usu-

al. Often the program guides him byasking "the next logical question-.

For example before the Lodwick andReichertz program will yield np anopinion concerning the likelihood ofbone tumor, it requires the resident

or fellow to make seventy observa-

tions and/or measurements on the

roentgenograms or patient. Even af-

ter the diagnostic programs have

evolved to the minimum number of

true discriminating factors, it seems

likely they will still be able to ask forobservations and measurements

which will help the beginner learn

medicine and the practitioner remem-

ber it.There are innumerable examples of

computer terminals and apparatusesdesigned to give the researcher a new

way to visualize his data. Multi-dimen-

sional plotting and display, symbolic

representation of data, conversationalmode of statistical analysis all offer

real potential for insightful analysis

of medical data from groups of pa-tients. It remains for us to learn touse modern display devices so as to

"give a new dimension" to the data

for the individual patient at the time

of his illness.

REFERENCES(1) Lindberg. D.A.B.: Electronic Retrieval of Clinical Data. Journal of Medical Education

Vol. 40 No. 8, (August) 1967, pp. 753-759.(2) Buck, C. ft., Jr., Reese, C. R., and Lindberg, D. A. B.: A General Technique for Com-

puter Processing of Coded Patient Diagnoses. Missouri Medicine, 83: 276-279; Passim.,(Apr.) 1988.

(3) Lindberg. D. A. 13.: The Computer and Medical Care. Sprin,gfield, Charles C. ThomasPublisher, 1968.

(4) Lindberg. D. A. B.: Collection, Evaluation, end Transmission of Hospital LaboratoryData. Methods of Information tn Vedicine, Vol. 6, No. 3 (July 1987), pp. 97-107.(5) Austin, C. J.: "The Medlars Profea at the National Library of Medicine." Library Resourcesand Technical Services, Vol. 9, No. 1 (Winter, 1965), pp. 94-99.

(0) Photographic and Computer Systems In Biomedical Information Handling, L. C. Kings-land, American Documentation Institute, 26th Annual Meeting, Chicago, Ill., October1983.

(7) Meyer, Thomas C., Departmcv of Postgraduate Medical Education, The University ofWisconsin. Personal communtzaion to the author.

(8) Stacy, R. %V., and %Vaxman, B., (Ed): Computers in Biomedical Research. Vol. 1. NewYork, Academic Press, 1965.

(9) Slack, W. V., Hicks, C. P., Reed, C. E., and Van Cura, L. j.: A Computer-based MedicalHistory System. New England Journal of Medicine, 273: 194-198, (Jan. 27) 1966.

(10) Caceres, C. A.: Electrocardiographic Analysis by a Computer System. A.M.A. Archivesof Internal Medicine, 3: 196-202,(Feb.) 1963.

56

11.e.

,

t.)

tf?'ay:

SUMMARYCOMPUTERS IN CLINICAL MEDICAL EDUCATION°

Panel: Donald A. B. Lindberg, M.D., Chairman, Robert Crecnes, M.D., Mitchell Schorow, JohnA. Starkweather, Ph.D., James Sweeney, Ph.D.

Dr. Robert Greenes described thepresent computer system at the Massa-chusetts General Hospital which utilizesa conversational mode for the collectionof information for the patient medicalrecord. This system usts a branching re-vonse approach to key5oard entries onthe part of the medical user. Both tele-type terminals and cathode ray tube ter-minals are employed. The latter includea special device produced at the Massa-chusetts General Hospital that consistsof a number of fine parallel wires cross-ing the face of the cathode ray tubescreen. When the medical user touchesone of the wires to indicate his choice ofa response displayed on the computerterminal, his body acts as capacitance toactivate a circuit whereby the computersenses his choice.

Dr. Greenes stressed both the needfor relatively carefully formated entriesin the patient record and the need forthe medical user to enter free form re-sponses on the computer terminal. Insome cases the standardized or formatedresponses are "fixed field" although thisis not always necessary. Processing ofthe totally free form entry is possible be-cause the conversational mode has al-ready contributed a considerable struc-ture to the data.

This system is being utilized in an ex-perimental mode on a regular basis forthree patient record data collection com-pilations. These arc the automated pa-

tient history, the conversational modesystem whereby the M. D. can answerthe results of his physical examination ofthe patient, and a system through whichthe radiologist is able to enter hisimpressions following interpretation ofroentgenograms.

Dr. Greenes, Dr. G. Octo Barnett, andtheir associates want to extend the ex-perimental system in order to investigatenew applications and new areas in thehospital setting. Dr. Greenes stressedthat one of the advantages of this typeof system, especially in the example ofthe radiology system, is the ability thesystem gives medical users to monitorthe thoroughness with which x-rays areexamined. This can play a definite rolein the educational process of residentsand young physicians studying radiolo-gy.

Dr. Greenes did not see any majorproblem in the existing hardware orcomputer soft ware. The major problemat the moment is the extreme amount oftime involved for the radiologist toproduce a dictionary, branching code,instructions and the thesaurus capabilityrequired by such a data collection sys-tem. It is possible that a radiologistcould spend the better part of a full yearin designing a system to handle only aportion (perhaps as small as fifteen percent) of the total dictionary and asso-ciated systems required for the field ofdiagnostic radiology. Consequently, the

°Presented by Donald A. B. Lindberg, M. D., Panel Chairman, at the closing plenary session ofthe conference.

57

Jô

limiting factor in this clinical data col-lection system is the amount of timepossible for clinical physicians to con-tribute to building ncw portions of thesystem.

Dr. James Sweeney described the his-tory of the Tulane information process-ing system as well as the present morcadvanced version blown as GIPSY. Hestressed the general problem of estimat-ing the true information content of datacollected through computer systems. Dr.Sweeney's system deals only with infor-mation which can be represented by al-phanumeric characters, and only infor-mation as it is perceived by the human"sensor." The major objective for indi-viduals designing computer systems isnot so much that the system need be atotal and perfect reflection of the realdata, but that it not lose any informationwhich has been perceived by the humanobserver. The system then becomes a re-pository for the observations of the hu-man being, and should provide him ca-phaility to inquire into this repository.In order for the systems to be at allmanageable and realistic in a medicalsetting, Dr. Sweeney felt it essential tostrip off from the total information ofthe hospital all that is not likely to haveessential worth in patient treatment andmedical teaching. Dr. Sweeney dis-dained the two extreme views: the viewthat all clinical information is exceeding-ly unreliable and "noisy" and the viewthat leads some people to wish to collectall information connected with patientcare. Dr. Sweeney characterized the lat-ter approach as the "pack rat principle"and pointed out that one should have afairly good idea of the use to which in-formation will be put before committingoneself to a computer data collectionsystem.

Dr. Sweeney stressed that the medicalrecord has very many functions. It isnnt, in his opinion, well organized at

58

present for purposes of sc

quiry. He also raised thwhether medical peopletraincd in the normal courseucation for conducting scicnics into information systenpresent paper medical re(

with computer repositoriesinformation, Dr. Sweeney felnew approaches and possiblmal education of medicalmight create a more useful t

At present, Dr. Sweeneying the reprogramming of h15formation processing systemwith the needs of I.B.M. 36Chardware. He expects to bea year and expects the systenerally usable by other investimedical and general scicSweeney's advance planningscheme to provide on-line investigators into their researe'using computer terminals. 1-1collection will remain for tthe paper form, utilizing certcoding principles developeyears by him and his grow.

Dr. Sweeney acknowledgeence of a problem in the exifacturer-supported progranguages. On the othcr hand,major problems to be still indefinition by medical persouses to which informationcareful systems design, and tty for innovative uses of comdeveloped in the future. Hethe inquiry capability he plaop has the possibility to givmension to the interphase Ihuman and the computer dahad developed interactive irbility into his system usingeration hardware. Under ttions, the response time inwas three to four minutes. Ihis. new system can reduce

to four seconds, and that response timesin the range of one second will probablybe required in order to obtain optimalinteraction between a research investi-gator and his data base.

Mitchell Schorow described the workthat Dr. George Miller and he arc doingin primarily addressing themselves tothe problem of educating medical per-sonnel for clinical problem solving. Thesystems with which they arc workingpermit them to create within the com-puter system the patient history database utilizing a single actual patient his-tory, a combination of actual patient his-tories, and whcre necessary, artificialconstructs. The computer system per-mits the student to "treat" the patient bypermitting him to ask the system for in-formation about the patient and his re-sponses utilizing virtually free form nat-ural language. For example, the systempresents to the student the informationthat he has a "patient who is a forty-nine year old white male with low backpain of two years duration, of increasingseverity for the past six months." Thestudent is then given the opportunity tomake diagnostic investigations andtreatment of thc computer "patient."The responses of the system are de-signed to contribute to the student'sgaining clinical acumen and good judg-ment.

Mr. Schorow and Dr. Miller arc utiliz-ing I.B.M. 1500 computer hardware andwriting in Coursewriter II language. Atpresent, they believe that their system iscapable of giving totally appropriate re-sponses to the student's free form ques-tions in about seventy per cent of theeases. This is in contrast to their beliefthat a minimum of ninety per cent ap-propriate responses are necessary in thelong run for the system to be a completesuccess. In fact, their system is still in anexperimental stage, although operatingin a gratifying manner.

59

Another major obstacle for these in-vestigators to overcome is the amount ofauthor time required in order to create agood student teaching program. At pres-ent, about forty hours of author time arerequired in order to produce one to twohours of student teaching material.Members of the panel were quick tonote that even the forty hours representsa very considerable improvement overthe amount of author time which wasrequired by other computer-aided in-struction systems. In many cases in thepast, one hundred to two hundred hourson the part of the author have beenneeded in other systems to create onehour of student teaching material.

Response time of the present Schorowsystem averages from four to five sec-onds, with a maximum response time ofthirteen seconds. Mr. Schorow felt thatany response in excess of eight secondsin such a system was totally unsatisfac-tory, since the delay left the student in-tolerably impatient. His feeling is thatwithin the next two years, the responsetimes of their system will have been re-duced to acceptable levels.

Mr. Schorow stressed the need tohave built-in objective evaluation criter-ia in the development of all such sys-tems. He did not feel that his own sys-tem is sufficiently far advanced at pres-ent to report the complete evaluation atthis time. His anecdotal experience withthe system encouraged him to believethat the system was quite acceptable, infact, exciting to medical student users.About one and one quarter hours are re-quired for each student to complete asingle clinical medical problem.

The problems presented to Mr. Rho-row in creating a system capable of ac-cepting totally free entries on the part ofthe student were described as seriousones. In order to accomplish this, the au-thors have utilized prepositional andpropositional calculus, and other ad-

vanced theoretical information handlingtechnics within the programming sys-tem. He seemed to feel that these tech-nics will ultimately prove adequate tothe problem.

Dr. John A. Starkweather describedhis professional obligations at the Uni-versity of California. These include theduty to improve education for both themedical student and medical profession-al community as well as nursing stu-dents and personnel, and other medicaland paramedical groups within the en-tire medical center complex. In order toaccomplish the task, hc described the di-vision of his central computing staff intogroups concerned with education, re-search, and patient care, and operatingas separate units. All three of thesegroups as well as the scientific users,utilize a single computer facility, in theform of an I.B.M. 360/50. As a conse-quence, Dr. Starkw e a th er utilizesOS/360, operating in multiple fixed par-titions, including tele-processing parti-tions as well as batch-processing parti-tions. A second smaller computer(360/20) is included as part of the sys-tem. The interphase between the twosystems being "HASP" input/outputspooling routine.

Dr. Starkweather stressed that it waspossible to serve this diversity of com-puter users with existing technics. Onthe other hand; he emphasized his feel-ing that ideal and optimal application ofa computer in medicine, especially com-puter-aided instruction systems, defi-nitely requires new programming lan-guages. These languages must inter-phase between the medical educatorand the various rigorous and complexrequirements of OS/360 and its JobControl Language. One such attempt in-volving Dr. Starkweather's new pro-gramming language system PILOT wasdescribed. PILOT is itself written inPLI programming language. The intent

60

of this system is to aid in teaching prob-lem solving to students and in teachingthe method of inquiry into patient prob-lems to medical personnel. The ultimateaid of the PILOT system is to permitthc medical educator to create his teach-ing routines with a minimal amount ofprogramming knowledge, and to permithim to do some file handling through re-motc computer terminals when the sys-tcm is finally completed. Dr. Stark-weather stressed his belief that thc pres-ent tools for inquiring into computersystems, for teaching of problem solving,and for simulation as a means of com-puter-aided instruction are still extreme-ly primitive.

Even so, computer systems have thepotential for giving appropriate re-sponses to relatively free inquiries onthe part of students. Dr. Starkweathcrfelt that at this very moment, the basicapproach of branching conversationalmode structurcd inquiry presented byDr. Robert Grecnes from the Massachu-setts General Hospital system probablyis closer to being an operational realitythan the free form entry systems. In es-sence, hc believed that the matrix ofquestions and responses used at Massa-chusetts General Hospital is presentlytotally realistic, but that the system hehimself is striving to create for handlingfree form inquiries offers a greater po-tential in the long run, and that this po-tential can ultimately be achieved.

Dr. Starkweather stressed. two addi-tional conclusions: one, that the facilityof a large computer to store vastamounts of information, whether patientcare information or didactic information,will in fact, make a major contribution tomedical teaching; two, the extreme im-portance in the long run of the comput-er's ability to provide the knowledge ofmany medical authorities to help onesingle student.

Dr. Starkweather presented in some

,

V..

t.

detail an application of his program-ming system in which a student canlearn to conduct a psychiatric interview.A characteristic of this system is theease with which the novice computeruser can answer the system and beginhis learnilg experience. Dr. Starkweath-er stressed, the fact that although thesystem is obviously a simulation experi-ence, the individuals using it seem to en-ter very seriously into the simulation.For example, individual users seem tobe compelled to quickly answer somesort of response to the computer termi-nal .questions, even though they wouldprefer to think over the answer and eventhough the question may be an awk-ward one. In this respect, they seem toforget that the computer terminal itselfis so utterly patient.

At present, Dr. Starkweather's majorprogyamming system is operating wellwithin the medical center and has goodlanguage handling capability, althoughit has some temporary limitations on thenumber of terminal units it can serve.The language is designed to aid the userby offering conversational mode assist-ance to him while his programming ef-forts are being conducted. The system isbased on a former similar system withwhich Dr. Starkweather had had experi-ence, namely Computest on the I.B.M.1620.

The major problems Dr. Starkweatherdescribed involve very serious limita-tions in present manufacturer-supportedprogramming language. He plans to cor-rect this problem by extending his owncompiler/interpreter programming sys-tem. He, the other members of the pan-el, and numerous members of the audi-ence were in agreement concerning thisissue: they all stressed their desire forfederal government and manufacturersupport for the development of new lan-guages for medicine and specifically formedical education. Dr. Starkweather

81

suggested four specifications for anysuch new language.

1. The language should offer ease ofentry to the computer novice.'In connec-tion with this, he stressed the extremelypoor quality and total unreadability ofexisting manufacturer produced pro-gramming manuals.

2. The language should offer case ofediting; that is, facility with which theauthor can make minor alterations in histext and instructions.

3. The language should offer ease oftranslation from similar or more primi-tive programming languages. This is ex-tremely important so that instructionalmaterial can be exchanged betweenmedical users in different institutions.

4. The new language should offcr theability to print for the user his sourcelevel instructions utilizing more thanone level of readability. This wouldmean that the individual relatively unfa-miliar with the program or even otherprogramming languages could request a"print out of the program" in a relativelyconversational, even verbose, form. Onthe other hand, the experienced individ-ual should be able to request a state-ment of the program in very much moreterse, compact, and efficient style.

A number of members of the audienceexpressed their agreement with the re-quirements Dr. Starkweather laid out aswell as the general inadequacy of pres-ent programming languages. One mem-ber suggested that with respect to medi-cal education, we were at a point analo-

gous to the days of scientific program-ming before the existence of FORTRANand other appropriate scientific high

level languages.There was considerable discussion at

different portions in the afternoon meet-ing from the audience concerning therecurrent theme: why are there notmore existing completed courses in med.

icine available through computer-aidedinstruction. All participants seemed inagreement:1. That there are no completed medical

courses.2. That even the most promising provi-

sional courses at the present willhave to undergo extensive medicaltesting and evaluation.

3. That an entire medical course, atleast in the traditional medicalschool sense, is unquestionably fartoo large a unit with which to be-gin attempting computer-aided in-struction. Many participants feltthat a single one hour lecturemight represent a realistic unit forthe medical educator to approachwith computer technics. Other par-ticipants, especially Dr. Greenes,preferred to approach a clinicalproblem rather than either a lec-ture or a course. The magnitude ofthe problem was hard to representbut he agreed that it might well beof the same order of magnitude asa single lecture.

There was considerable discussionaimed at explaining the present essen-tially primitive state of computer-aidedinstruction in medicine. Many explana-tions were offered. A popular oneseemed to be the belief that the prob-lems in medicine were essentially verymuch more complex than the problemspresented in teaching undergraduatesubjects of a seemingly more quantita-tive and more rote nature. Exampleswere offered: statistics, languages,grammar school spelling and arithmetic,and college physics. The latter examplewas challenged by one of the panelmembers who suggested that it had costmore than $200,000 to program one col-lege physics course, and that the finalresult had not been totally acceptable tothe teaching department involved. Con-sequently, one was left with the feeling

that the participants generally agrecdcomputer-aided instruction could beutilizied in areas where the teachingunit was sufficiently small, and thematerial to be covered was sufficientlywell defined; but tbat in fact, neither ofthese conditions had yet been identifiedin a truly medical setting.

As an extension of the above, therewas considerable discussion from boththe audience and the panel concerningthe question of new roles for computersin medical education. Virtually everyonewas willing to concede that the use ofcomputers to compile and give inquirycapability to the patient record keepingsystem probably provided a suitable as-set ia a medical education setting. Therewas, however, very little enthusiasmfrom the audience to pursue this prob-lem during the meeting. There seemedto be willingness to believe that this en-deavor was worthwhile, especially if thequestions asked such a computer database were aimed at an epidemiologidalor population level, and especially ifepidemiologists were able to play a rolein teaching students to ask such ques-tions.

In a similar way, there was a greatwillingness on the part of the partici-pants to accept the contention that ov-ertly ins tructiona 1, traditionalcomputer-aided instruction technicscould take their place in the long run asa valid portion of the medical educationarmamentorium. Even so, most peoplewere willing to admit that accomplish-ments in this sphere were presentlyquite limited and the technics essentiallystill primitive. For this reason, there wasa very general consensus that the profes-sion should continue vigorously to seeknew irriaginative applications for com-puters in teaching. This might take theultimate form of diagnostic games, com-puters as an aid in logic and reasoning,computer simulation of mental and/or

physical processes or as yet totally un-discovered applications. In order to pur-sue this course of action, it was agreedthat one should make every attempt toinclude objective evaluation criteria intosuch experiments. It was the belief ofthe section chairman, to which he gar-nered surprisingly little support, that theproper application of cost effectivenesscriteria cannot be made to the experi-mental systems unless one carefullyspecifies the states of developmentthrough which such systems need topass. In this sense, he was urging theapplication of staged development tech-nics utilized in engineering; for example,creation of specifications and require-ments, pilot model, prototype model,test model, production model, and finalevaluation. In defense of this contention,it should be noted that there was onlyone engineer in this section meeting. Hewas an industrial engineer and con-curred with the desirability of stagingexperiments. -

A member of the audience asked thegroup for general recommendations fora university administrator in planningfor computer terminal capability in thedesign of a new campus structure. Theconsensus was that:I. The school should make provision for

possible coaxial cable connectionsbetween the new building and thecampus computer site (probably

this would mean a tunnel).2. With the new building, suitable tele-

phone line conduit should be pro-vided to each officc and laboratory.

3. False or "deaf" ceilings should beused.

4. Air conditioning should be provided.5. No cables or electrical connections

should be buried in concrete, norshould "channels" in concrete beused.

These were general recommendations.The group agreed that provision mightbe made for a small peripheral comput-er in the new building if a large number( more than eight or twelve) computerterminals were planned. A small com-puter would also probably he neededwithin the building if complex physio-logical simulations were planned (suchas the automated anesthesia dummy de-scribed by Dr. Abrahamson).

Inquiry was made of the group for apublished estimate of national, state, orregional manpower needs with respectto medical computer personnel. Primari-ly the question here was the need fordevelopment of medically oriented com-putcr systems and operators, computerprogrammers, and information scientists.All those asked felt most such personswere in critically short supply but couldneither document this need nor identifyinstitutions at which. such persons arebeing trained.

In summary, the participants were enthusiastic about the potential of contri-butions of computers to medical education. They seemed to feel generally thatthis field is presently in its infancy; that it should be explored in an openly exper-imental framework. A major limitation to progress in the field at present is theexhorbitant amount of author/clinician time required for the creation of even ex-perimental systems. The participants seemed to feel that this problem was bestattacked by the creation of new computer programming languages, pre-lan-guages and/or systems, of a sort to support conversational mode interaction be-tween the relatively inexperienced user of computers and the necessarily largeand complex computing systems which will be required.

,63:0 4

COMPUTERS IN CONTINUINGMEDICAL EDUCATION

GEORGE E. MILLEFI, M.D.°

I will begin with a disclaimer. Myknowledge of computers is best de-scribed as rudimentary, and if I speakwith any semblance of confidence aboutbits and core, of CRT and 360-50, it isbecause my associates have pro-grammed me with some care to respondappropriately to natural language input.Having made such a confession I mustalso acknowledge that my justificationfor accepting an invitation to appear onthis program is based upon a conclusionthat the key word in context of theconference title is not computers but ed-ucation, a field in which I feel some-what more comfortable.

For in fact, the computer is neithermore nor less than a tool, neither laud-able nor objectionable except in terms ofthe task it is to perform. And when thattask is continuing education then theproblem is formidable, for debatesabout the perplexing and frustrating is-sues in this field grow ever more shrill.Of only one thing do we seem certain:the pressure is mounting for the univer-sities and others to do somethilig, prefer-ably something new and hopefully withnew tools as well. Among the new toolsto which we are turning, the computer isnearing the crest of a wave of enthu-siasm that was only recently vacated byprogrammed instruction. Discussion ofcomputer assisted instruction is becom-ing a babel in the land, and talk of net-works and compatibilities and software

can be heard almost everywhere. But inthe language of communication theoriststhe educator must ask whether all thissound represents a message or merelynoise.

I am personally inclined toward theview that there is not only an importantbut a critical message to be heard, butone cannot deny that the noisemakersare also plentiful among us. These arethe ones who seem to be saying that thisdazzling new medium is the messageand like the medical educators beforethem who made the same claims aboutfilms and television and programmed in-struction they divert us from the centralquestion of educational purpose to beserVed by such exciting tools.

This cautionary comment must have afamiliar ring, but it is now touched witha note of urgency. For the widespreadrestlessness about improving means ofbringing an ever growing body of medi-cal information more rapidly to practi-tioners is accompanied by a mountingfaith that the incredible informationstorage and retrieval capacity of thecomputer can accomplish this allegedlycriticA task. While it would be foolhar-dy to deny the accelerating pace ofknowledge growth, it is at least defensi-ble to raise the question of whether theproblem of continuing education is less-ened or magnified when a practitionerseeking information gets a MEDLARSprintout of eighty citations on the topic

°Director of Research in Medical Education, University of Illinois College of Medicine

64

-0.0

about which he has inquired. The tech-nical achievement of such a production isstaggering, but the educational impactrhay leave more than a little to be de-sired.

The educational issue then is not com-puter capacity and versatility, but ratherwhat kind of information shall be stored,in what manner it should be processedand retrieved, in order to serve what ed-ucational objectives. In my own trou-bled search for an answer to this ques-tion I have come, on April 4, 1968, to apoint of identifying four functions towhich the computer seems both readyand able to make a contribution to con-tinuing medical education. I share themwith you not in any confidence that theywill survive your scrutiny, but in thehope that they may serve as organizersfor our discussions this afternoon.

The first, to which I have already re-ferred in passing and which will bemore fully developed by Dr. Cummings,is the library function. I find it difficultto believe, as some seem to imply, thatthe library of the future will be a collec-tion of discs and tapes and punch cardsfrom which words will flow as a steadyelectronic stream in response to thecommand of a discriminating client. Butthe evidence is persuasive that there aremore efficient means of searching for aninformational bit than poring througha card catalogue or the Index Medicus.The computerized search function, withretrieval of selected abstracts or rapidreproduction of full texts will surely bean important contribution to the physi-cian who has a specific question forwhich a specific and documented an-swer is available. Aside from the techni-cal problem this poses, there is an edu-cational problem of retrieval, of match-ing level of inquiry with level of re-sponse, and not overwhelming the in-quirer with more information about pen-guins than he cares to have. I suspect

65

that this library function will be ofgreater significance in the continuingeducation of academicians (who seemto have a high level of tolerance for ir-relevant information) than for practi-tioners but that is a matter of specula-tion.

The fact remains, however, that thisfunction will be useful in fulfillment ofonly one objective of continuing educa-tion, and one which at least some of usbelieve to be of secondary rather thanprimary importance: that of providinginformation. It does not guide learning,but responds to practitioner-generatedinquiry which can as easily be a majoreffort to find small things as essentialstudy of critical matters. Such randomactivity must sometimes be best de-scribed by Dorothy Parker's phrase,"There is less here than meets the eye."

Secondly the computer can serve aconsultative function. It has been saidby some, although without benefit ofany discernible evidence, that the bestcontinuing medical education occurswhen a consultant and practitioner worktogether in solving a perplexing clinicalproblem. To the extent that this is true,the role of the consultant-teacher wouldappear to be primarily that of bringing aspecial body of information, or a partic-ular skill in applying both familiar andnew knowledge, to the problem at hand.It is such a function that seems to beserved by some of the imaginative com-puter applications that have alreadybeen introduced into clinical practice insome parts of the country. For example,the computer analysis of electro-cardio-grams which not only frees a physician-interpreter from routine tasks but alsofocuses his attention upon elements ofthe tracing that require special consider-ation; the clinical diagnostic systemwhich guides a practitioner's inquiryand from a vast memory core can sug-gest diagnostic probabilities, as well as

possibilities that have escaped his atten-tion; the rapid calculation of radiothera-peutic dosage which incorporates con-sideration of many variables that mightotherwise have been neglected or omit-ted; the continuous monitoring and syn-thesis of physiologic data from patientsin coronary care units, patients whosecourse may be modified by early recog-nition of impending complications; therapid analysis and weighting of histori-cal items in screening examinations,with probability projections that allow apractitioner to focus his attention uponthe most significant matters in his en-counter with an individual patient. Allof these may provide substantial andsignificant assistance to a physician inthe fulfillment of his professional role,and thus contribute to solution of thecritical manpower shortage which besetsus. But the question that must occupyour attention here is whether thesethings, or indeed whether more conven-tional consultative exchanges, are educa-tional in the sense that they extend thelearner's ability to deal more efficientlyor more .effectively with comparableproblems he encounters in the future, orin another setting. It is surely possiblethat they may do so, and I am preparedto accept this as a potential, if not a de-monstrably achieved outcome. But atthe same time it seems important to notethat a crutch may lead to dependence aswell as to independence. Nor does theuse of a more efficient kind of crutchnecessarily influence the nature of theoutcome. There would probably be nodispute, with the view that electronicmeasurement of the hemoglobin level ismore precise than the Tahlquist method,but there might be considerable ques-tion whether the introduction of accura-cy has led practitioners to more rationalpractices in the treatment of anemia.

If the consultation function is to beuseful as an educational instrument it

66

must probably be accompanied by somesystematic review of what has beenlearned, not merely what diagnostic ortherapeutic goals have been achieved.

Third is the tutorial function, perhapsthe most familiar of the species includedunder the genus computer assisted in-struction. It is this form that many vi-sionary directors of continuing educa-tion programs have in mind as they pic-ture a computer network for a city, a re-gion, a state, or in the case of EDU-COM, a nation. Such a mode lends itselfto discrete course packaging, and to sys-tematic consideration of highly impor-tant clinical content and concepts. Withthe complex branching capabilitieswhich even a simple computer offers,the instruction may proceed not only atan individual learner's pace, but maybypass informational bits with which heis already familiar. Unlike the book,which also offers these advantages overclassroom instruction, the tutorial formof computer instruction forces activeparticipation in the learning process andsome assurance that the necessary con-tent has been mastered to a predeter-mined level before proceeding to thenext stage. And the recitation, whichmost such programs require, has the ad-vantage of being personal and private sothat the revelation of ignorance carriesno social threat, unless the learner per-sonalizes the computer-tutor and tries tobeat the transistors in much the samefashion that students have for so manyyears tried to beat the living teacher.

Recognizing these several advantages,which reflect the incorporation into edu-cational technology of substantiatedtheoretical concepts of learning, thereare still many who cry out in despairthat such mechanical and dispassionateinstruction might displace the warm andpersonal interaction between live teach-er and live student. I suspect that mostsuch complaints come from teachers plc-

turing what they think they do with stu-dents rather than from any accurate de-

piction of what occurs in lectures,conferences, even seminars, in hundredsof continuing education programs all

over the nation. Is listening with onehundred others to an expert talk (withthe opportunity to ask questions at theend of the hour if there is time left)really personal instruction, while engag-ing in active individual analysis, in-

terpretation and discussion of informa-tion with a carefully programmed com-puter really impersonal? And for thosewho insist that such an exchange is mere

training, not real teaching, I would notethe comment by B. F. Skinner in a re-cent Science paper that when we knowwhat we are doing it is called training,and when we don't it is called teaching.

Despite the advantages of plannedcoverage, carefully specified learningobjectives, active learner participationand immediate feedback, the tutorial

mode of computer assisted instruction isstill focused upon information which hasbeen selected and processed by a teach-er, and whose acquisition is assumed to

be accompanied by transfer into profes-sional practice. It is this assumption thatmust be most seriously questioned, for

practitioners have long known, for ex-ample, that bacteriologic identificationof an organism should precede institu-tion of antibiotic therapy or, at the veryleast, confirm the choice after the fact,but even casual inspection of the realityreveals the frequency with which thislogical pattern is ignored. Such discrep-ancy between knowing and doing is the

most disquieting problem of all for

thoughtful leaders in continuing educa-

tion.It is for this reason that I am particu-

larly attracted to the fourth educationalfunction the computer may play the

simulation function. In order to avoidany potential argument let it' be openly

67

:.1

acknowledged at once that simulation is

not reality. But at the same time let theclaim be underscored that for some in-struction it may be more useful thanreality. For in simulation it is possible toextract a critical and planned segment ofan often complex and unpredictableworld; to control the evolution or se-quence of important variables; to pro-vide learners with opportunity for inde-pendent action without doing harm toeither subject or object; .and throughthis dynamic mechanism to provideimmediate feedback of consequences ofan intervention, not merely informationabout what might transpire. It is an in-structional form that focuses upon userather than mere possession of knowl-edge.

There are already exciting develop-ments in computer simulation, rangingfrom relatively simple to dismayinglycomplex. Spivack's arrhythmia trainer,one of the simpler applications, creates asituation in which sudden changes inelectrocardiographic pattern displayedon an oscilloscope demand prompt andproper application of electrical stimulito a synthetic torso in order to reverse,for example, ventricular fibrillation or toinitiate appropriate pacing of a poten-tially disastrous heart block. A morecomplex mode is Sim I, Stephen Abra-hamsons' life-size dummy whose respira-tory activity, pulse rate, blood pressure,skin color and pupillary size, among oth-er things, are under computer controland respond to drug administration aswell as to other interventions used bythe anesthesiologist in managing pa-tients at surgery. Using this model, atrainee can be faced with a variety ofsituations he will shortly encounter inthe operating room, and allowed to dealwith them repeatedly, without risk toreal patients, and with immediate feed-back about the effect of his judgments,and his actions.

Another program is John Starkweath-er's simulation of the clinical interviewin which a student may play either therole of physician or of patient, while thecomputer carries out die other part ofthe dialogue.

The method with which I am most fa-miliar, however, is that under develop-ment in our own unit under the direc-tion of Mitchell Schorow. This hasgrown out of the printed patient man-agement problem, the erasure examina-tion which has become one of the morefamiliar simulation methods in currentmedical education. Here the learner isprovided with the bare bones of a prob-lem usually little more than the pa-tient's presenting complaints. Selectionof general problem-solving strategiesand specific management tactics mustthen be independently generated usingnatural language input, and with re-sponses provided through typed output,visual display ( such as x-ray, electrocar-diogram, biopsy specimen), or sound( such as heart sounds, lung sounds, andsegments of spoken historical informa-tion that provide non-verbal as well asverbal cues ). Management interventionsare followed by alteration in clinicalcourse, such that a learner may deal suc-cessfully with a problem, inadvertentlyinduce complications with which hemust then deal, or lose the patient to an-other physician or to the pathologist.But Lazarus-like, this patient can risefrom the dead, and if necessary sufferthe same fate again and again.

While such an encounter is demon-strably exciting, we must certainly avoidthe comfortable trap of equating excite-ment with learning. They may be asso-ciated, but it is not necessarily so. Thecarefully created simulation, however,does provide an unusual opportunity todetermine the means by which a learnerattacks a problem, the types of evidencehe accumulates before taking action, the

68

0

relative weight he assigns to evidence ofdifferent kinds, the extent to which ef-fects of early interventions influence hissubsequent actions, and his manner ofdealing with deviations from anticipatedoutcomes. If the sampling of this behav-ior is sufficiently large it may even allowsome generalization about his character-istic pattern of performance which,when cast against the judgments of acriterion group, can focus attentionsharply upon the strengths and weak-nesses of his knowledge and his methodsof utilizing knowledge. In short, it pro-vides diagnostic data which may then beused to prescribe educational therapyspecifically selected to correct his per-formance deficits therapy that may bein the form of books, journal articles,courses or tutoring.

It is such a posture that has been as-sumed by the EDUCOM Task Force onContinuing Education that education-al therapy abounds, but diagnosis ofneed is in very short supply. With theEducational Testing Service and re-search groups in medicine, engineeringand education a proposal has been de-veloped to study use of the computer,among other methods, for the purposeof developing such educational diagno-sis, and not merely the shotgun therapywhich seems to characterize so much ofcontinuing education. It is the hope ofthis study that a library of simulationsdesigned to provide the basis for diag-nosis will not only direct attention to in-dividual and group needs but will alsoheighten the awareness that learning be-gins with recognizing a question, nothaving an answer, with the process ofseeking and not alone with some prod-uct that others have found.

Such a shift in educational strategywill not be easy to achieve, for ft- achershave long managed to divert a variety ofinstructional methods to their ownneeds. There is no reason to believe that

a computer will accomplish a magicaltransformation of teacher-centered in-stuction to student-centered learning.But it does create an opportunity for in-dividualization of learning that one canhope will be grasped and exploited tothe full.

The developmental stage of any newmethod is certainly not the time to raisequestions about cost, yet it is an issuethat must ultimately be faced both real-istically and quantitatively. The currentenchantment with computers will notlast forever, and when the enormousbills for programmers and analysts, longlines and equipment begin to seem bur-densome, we had better be prepared todefend with something more solid thanenthusiasm our view that computer-aidedlearning for practicing professionals is

not only effective but competitive. As

69

the investment in hardware and soft-ware goes steadily upward, I wouldplead for the diversion of some signifi-cant segment of that investment, in ev-ery developmental program, to system-atic study of use ( once the bloom ofnovelty has faded), of cost ( once thedevelopmental price has been paid), butmost importantly of effectiveness. And itis here that we will require a measure ofhumility that has not always character-ized the physicians who conceive pro-grams of continuing education, or themedia masters who have nourished andsupported them. Our greatest need is aspirit of inquiry not about computers

(in continuing education), but aboutcontinuing education (with computers).

And oh that shrill note I shall retire tothe wings.

SUMMARYCOMPUTERS IN CONTINUING MEDICAL EDUCATION°

Panel: George E. Miller, M.D., Chairman, William T. Blessum, M.D., Edward N. Brandt, Jr.,M.D., Ph.D., Gerald Fscovitz, M.D., Alexander Schmidt, M.D.

The panel and wdience were chargedby its chairman with development ofdiscussion and recommendations whichcould guide those interested in workingwith computers in Continuing MedicalEducation and those who may have re-sponsibility for funding such work.Members of the audience introducedthemselves and their major interests andquestions. It was found that althoughthe audience consisted of a diversity ofbackgrounds and interests, their ques-tions were very similar. The audiencewas actively involved in the discussionand this report includes their contribu-tions as well as those of the panelists.

The first major consideration was therelationship of computer based instruc-tional systems to other computer basedhealth care systems. Dr. William Bles-sum of the California College of Medi-cine pointed out that he had been activein attempting to stimulate interest incomputer based systems by pilot appli-cations for medical students. The com-puter group has been active for aboutnine months and there is interest in de-vf.:lopment of a system which utilizescomputer-assisted instruction to themaximum extent possible. At the presenttime it is difficult to decide on which in-structional system to use since a numberof significant ones are available. Fur-thermore, there are a wide variety ofterminals and philosophies of design forthe operating system in which comput-

er-assisted instruction takes place. It is acontinuing problem to determine howbest to take advantage of the variousprograms available because of the widevariety of hardware configuratioirs. Dr.Blessum challenged those engaged incomputer-assisted instruction to realizethat it is in an embryonic state of devel-opment and must be approached in aninvestigative atmosphere. He empha-sized the need to be as compatible aspossible and yet to realize that there is adanger in being compatible too earlysince this may restrict innovative ap-proaches. Therefore, he emphasized theneed for generation of a master plan forall applications of computers in medi-cine so that library systems, medical in-formation systems, computer-aided in-struction, etc. will all be components ofthe same general system. The advan-tages and disadvantages of such an ap-proach weze discussed. Specifically, itwas pointed out that building a comput-er based instructional system as a partof a total health information systemcould provide a means for stimulatingphysicians to accept the computer as aninstructional tool and to use it. The prin-cipal problem to be faced by developersof computer-aided instruction is how toget this form of instruction to physi-cians. Should terminals be put in doc-tor's offices or can they take advantageof this through other means? It was feltthat if computer-aided instruction were

Presented in closing plenary session of the conference

70

0 9

by Edward N. Brandt, Jr., M.D., Ph.D.

part of a total health information sys-tem, one could more readily justify pro-viding terminals for physician's use sinceit would be possible to have at his dis-posal many functions, such as summa-ries of clinical data, diagnostic assistance,information retrieval, etc. that would beof value to him. Furthermore, this wouldprovide a means for obtaining feedbackfrom the physician so that he could helpdefine and contribute to the objectivesof his own continuing education fortranslation by educators and computerscientists. The question as to whetherphysicians would use terminals for anypurpose was discussed. It was allegedthat medical students do accept themPnd if a computer based health informa-don system were able to play a consulta-tive role and lead to improved patientcare, physicians would make use of it.However, much discussion centeredaround the physicians acceptance of thecomputer especially as a part of his con-tinuing education. In summary, there isvirtually no information available as tophysician acceptance and careful studyis needed. A computer based health caresystem could provide high quality careof patients since it would have the ad-vantage of having necessary data avail-able and the results of group experi-ences to assist in judgments. However,in the development of such a systemthere is the need to separate judgmentsfrom facts and to provide for grouprather than individual judgments.

Another advantage to introducingCAI by this means is that it would havethe virtue of peer group pressure on thephysician to take advantage of a verypotent tool. However, physicians wit!make use of it only if the computerbased system improves the quality oftheir activities and if the system is welldesigned and meets their needs. All ofthis discussion was based upon accept-ance of the concept that the purpose of

continuing medical education is to im-prove the quality of patient care. Thegroup was also reminded that knowl-edge of what constitutes good care anduse of such knowledge are not necessari-ly the same.

The disadvantages of building theCAI system as a part of a health infor-mation system were also discussed. Inthe first place, such a system may be-come a "crutch" for the physician andthereby, defeat thinking and learning.Furthermore, if the system provides forpunitive action or enforcement of partic-ular ideas, this may in itself defeat itsacceptance and lead to the developmentof poor attitudes towards its use. How-ever, such a system permits continuingeducation to be indirect and thereby notrequire the physician to spend blocks oftime in his education and this is desira-ble. Furthermore, the material presentedwill be relevant to the physician's activi-ties and this has the greatest chance ofinducing behavioral changes.

The discussion then turned to consid-eration of the objectives of a computer(based continuing education program.Quality patient care was consiclered to

be the most important objective of con-tinuing medical education, but relativelylittle is known about what motivatesphysicians to obtain further education.Studies should be carefully and system-atically designed to determine what theobjectives of continuing educationshould be. A very important point wasmade that getting information to physi-

cians was not as important as inducing

them to make use of this information.The material presented should be rele-vant to his own goals and therefore it isnecessary to determine what these goalsare and to translate them into meaningfuleducational strategies. The evaluation of

any instructional system must dependupon the realization of goals and there-

71

fore it is necessary to set objectives inadvance.

Other questions that need careful andsystematic study were also discussed.For example, does a computer based in-structional system actually inhibit learn-ing or perhaps better stated, under whatconditions does such a system inhibitlearning? If, for example, it were usedby the physician as a decision makingtool, it might influence the physician tonot continue his educat!on. Further-more, its use as a policing device wouldlead to rejection rather than stimulation.

Is it better to introduce computerbased instruction into medical schoolsand then continue with these studentsthan to try to expose physicians whohave developed study habits over yearsof training? The general problem of in-fluencing medical students and physi-cians to change their habits of study isfaced by educators interested in CAIsince it will not be used if it does notprove to be a more efficient and success-ful method. CAI does have the virtuepotentially of providing individuallybased instruction and thereby of beingan effective self study tool. However,the realization of this depends upon ac-cess to medical information on a veryside scale, hence, the apparent desirabil-ity of making it a part of a larger infor-mation system.

Discussion then turned to the role ofcomputers in continuing medical educa-tion. The tutorial approach may not beindicated in continuing medical educa-tion due to the potential lack of rele-vance to the physician's concern. Again,it was emphasized that the immediateproblem is not one of getting informa-tion to the physicians, but of stimulatingbehavioral changes based upon that in-formation. The tutorial approach is mosteffective in the presentation of factualinformation to be learned and hence by

its very nature is fairly rigid in design.Flexibility to permit the student to se-lect the material or portion thereof thathe is to learn is very desirable. More so-phisticated diagnostic methods to assisthim in this selection would be of benefitbut at this time such methods have notbeen developed. Computer simulation ofclinical situations that are revelant is oneapproach to get the physician involvedand to make it important to his ownpractice. Such situations permit the phy-sician to develop a solution with the as-sistance of the computer as a tutor. Ex-periments involving simulation of clini-cal situations are underway, hence, moreinformation about this aspect will beforthcoming. Again, the consultativerole of providing information and assist-ance in decision making by means oftie-up with a health information systemwas also pointed out. This provides thephysician with access to empirical aswell as theoretical information and ex-periences.

The consensus of the panel was thatthe goals of continuing medical educa-tion are to develop a set of attitudes to-wards: (a) use of records in the broadsense as a basis for judgments; and (b)appropriate medical care.

Therefore, it is recommended thatthere be developed carefully and sys-tematically planned investigations into:(a) what should be the specific objec-tives of continuing medical education;(b) will instruction developed from acomputcr based health information sys-tem enhance or inhibit learning; (c)how will physicians use such systems;(d) will such systems lead to increasedmotivation for further education; (e)how does one develop compatibility ofexisting systems; and (f) what are thepotential uses of computer-aided in-struction in continuing medical educa-tion? It was stressed that work to devel-

72

7 1

op a greater diversity of computer basedinstruction must continue and be evalu-ated before the role of the computer incontinuing medical education can be de-

-At

fined. However, such developmentshould not ignore, but should attempt togather information to answer the abovequestions.

AUTOMATION IN MEDICAL LIBRARIES*MARTIN M. CUMMINGS, M.D..1- AND RALPH A. SIMMONS$

In science, read, by preference, the newest work; in literature, the oldest.Edward Bulwcr-Lytton Hints on Mental Culture.

THE CHALLENGE

An important purpose of medical edu-cation is the transfer of skills, knowl-edge, and information from a variety ofsources through a variety of media tothe student and practitioner. A most im-portant resource in this process is themedical library. We have become in-creasingly conscious of the relative rolesof other media so important in the com-munication processes, including the in-formation transfer that takes placethrough personal discussions with col-leagues, attendance at professional lec-tures, and the use of the newer audioand audio-visual materials includingtelevision. But the literature the writ-ten word remains as a principal vehi-cle through which scientific concepts,facts, ideas and observations are trans-ferred.

The growth of the medical literaturehas almost doubled over the past dec-ade, as has been true in practically allscientific fields. The literature itselfabounds with articles discussing the ex-ponential rate of growth of scientificperiodicals and writings. The order ofmagnitude of this "explosion" has beenfrequently debated with estimates ofscientific periodicals being published to-day ranging from 35,000 to over 100,000.

It has been estimated that there aremore than 18,000 serial publications inthe biomedical field alone, with anequal number of medical books andmonographs generated throughout theworld annually. The growth rate hasbeen described by Derek de Solla Priceof Yale University. "The most remark-able conclusion obtained from the data. . . is that the number of journals hasgrown exponentially rather tban linear-ly. Instead of thcre being just so manynew periodicals per year, the numberhas doubled every so many years. Theconstant involved is actually about fif-teen years for a doubling . . ." ( 1 )Vannevar Bush wrote as early as July1945. . . "There is a growing mountainof research. But there is increased wi-dence that we are being bogged clowntoday as specialization extends. The in-vestigator is staggered by the findingsand conclusions of thousands of otherworkers conclusions which he cannotfind time to grasp, much less remember,as they appear?' (2)

Our society looks to the university asthe instrument of primary responsibilityfor the creation of knowledge throughresearch, its dissemination throughteaching and its application throughservice. The centers for education in thehealth professions have been developed

°This paper was prepared by Dr. Cummings and Mr. Simmons as an extension of the subject,"Computers in Medical Libraries," presented by Dr. Cummings at the conference.

tDirector, National Library of Medicine.:Head, Information and Computer Science Staff, Research and Development program, NationalLibrary of Medicine.

74

within this university framework to ac-complish this task with respect toknowledge in health-related matters. So-ciety originally looked to the libraries asinstruments to preserve recorded knowl-edge at a time when the corpus wassmall and manageable. Today they havetaken as an equally important functionthe organization and retrieval of neededinformation from the present enormousbody of knowledge. Thus the libraryrepresents a central learning resourceserving our universities.

Health science libraries are faced withan ever increasing challenge. This is notdue solely to the rapid growth in themedical and related literature; the wid-ening range of users and their accompa-nying diversity of requirements com-pound the problem. The list of usersmay include undergraduate students,graduate students, postdoctoral fellows,interns and residents, members of theteaching faculty in the basic sciences,clinical faculty members, practitionersin the health center or its environs, med-ical researchers, dentists, veterinarians.pharmacists, nurses, and other alliedhealth personnel. It may also includeother regional and hospital clients,health departments, welfare agencies,local industries, the county medical so-ciety, miscellaneous lay persons, attor-neys and others. The basic homogeneityof the principal users of a medical li-brary remains as an asset to the medicallibrarian in providing required servicesbut as this list suggests, that advantageis no longer so pronounced. The impor-tance of interdisciplinary application ofknowledge and research will continue torequire expanding services and capabili-ties of the medical library.

In addition to the quantitativechanges of information and users facingthe medical libraries, there have beenimportant qualitative changes in medi-cal education. Curricula have broadened

75

in scope and depth. New subject areashave been incorporated as knowledgeincreases. Undergraduates are more in-volved in research. The student mustabsorb more of the total body of scienceand then be responsible for keeping upwith new knowledge. It is no longer themedical undergraduate alone who mustbe educated, but the practicing physi-cian. B. V. Dryer in the article "LifetimeLearning for Physicians" has stated that. . . "The gap between scientific knowl-edge and application grows wider eachyear because of at least four reasons:the rapid advances of research; the mal-distribution of opportunqies for contin-uing medical education; educational in-adequacies even in those places whereopportunities do exist; and patterns ofeducational organization and dissemina-tion of knowledge which are not effi-cient in terms of the physicians students'needs." (3)

The opportunity to prepare studentswith the needed attitudes to meet thelater demands of continuing educationmust not be lost. The things whichyoung medical students learn in theirformal educational courses set the direc-tion of their future lives as physiciansand scientists. Harvey Cushing, who onmany occasions laointed out the impor-tance of the medical library for supportof medical educaticn, research and prac-tice, commented on the importance ofteaching students the value of their li-brary as follows: "To inoculate a doctorwith the library-habit he must be caughtyoung and here, as I see it, you have anexcellent opportunity in being nearer toa growing centre of medical education.. .

If you are to infect the young withthe reading habit, you must set the trapfor them, so baited that they will walkinto it unawares. Books must be madeaccessible. It is someone's business inevery medical school to teach laboratory

methods to the students but it is noone's particular business to teach themhow to use medical literature. . . ." (4)

The subject of the need for continuingeducation should not be left withoutsome reference to the place the hospitalis assuming and the role of its library inthis process. The hospital represents thelocation where the professional healthpersonnel can bring scientific knowl-edge to bear on a problem. Within thelast four decades, additional educationalprograms have been started within thehospital as it represents an unique envi-ronment where certain kinds of post-graduate education can take place. It isclear that the hospital must take increas-ing responsibility for education at alllevels. When the hospital is thus viewedas a scholarly institution involved in re-search and education as well as patientcare, the need for effective access tomedical and related scientific knowl-edge is self-evident. The dilemma is thatthe requirement for the availability ofthe scholarly record is so broad that nosingle hospital library has the resourcesto properly respond.

The problems of scientific communi-cation confronting medical librarians ap-pear to approach the insurmountable;quantity of information rapidly growing,quality of education and research im-proving, information user groups ex-panding, interdisciplinary requirementsenlarging, and education throughout aprofessional's career demanding support.It seems likely that the health scienceslibraries will have to take on a new iden-tity and function to meet this challenge.

Libraries are even now rapidly turn-ing to the electronic technology to re-spond to society's demands. The use ofadvanced data processing equipment,including computers, coupled with theresearch and techniques Geveloped andbeing extended in the emerging field ofinformation sciences holds great promise

76

for the solution of the biomedical com-munications dilemma. There is a warn-ing, however, that must be sounded be-fore exploring the applications and po-tentials of these new technologies. Wil-liam Os ler, on stressing the importanceof organizing and extending smaller li-braries in medical societies or hospitalsfor continuing education of physicians,wisely pointed out, "It is much simplerto buy books than to read them, andeasier to read them than to absorb theircontents." (5) In a similar view, it ismuch easier to introduce computers intolibraries than it is to design systems fortheir productive and efficient use.

Historical Development of LibraryAutomation

The mechanization of data processing,including both what is known as EAM(Electronic Accounting Machinery) andthe newer computer equipment, has anintricate and somewhat elusive history.As early as 1780 Joseph Marie Jacquardinvented a machine in which perforatedcards controlled the pattern of hooksand needles for use in weaving; theEnglish mathematician, Charles Bab-bage, conceived the idca of an automat-ic digital computer about 1835. Hecalled his proposed machine an "analyti-cal engine", but was unable to completea working model before he died in 1871.(6) It is generally acknowledged thatthe first operational device using perfo-rated cards for tabulation and statisticalpurposes was developed in the 1870's,and represents the real genesis of thepresent multi-million-dollar data pro-cessing industry.

The invention of the electrical tabu-lating machine using punched cards isgenerally attributed to Herman Holler-ith, a statistician, who was working atthat time for the United States CensusOffice. But from his own records, andthose of others, it is clear that it was a

physician who first suggested to Holler-ith that an electrical machine utilizingpunched cards should be developed foruse in the 1880 census. This physicianwas Dr. John Shaw Billings, the princi-pal figure in the history of the present-day National Library of Medicine, serv-ing as its first director from 1865 to1895, and unquestionably the pre-eminent figure in the entire field ofmedical librarianship.

Serving as a consultant to the CensusOffice, Billings suggested in 1880 "thatthe various statistical data of the livingand the decedent might be recorded ona single card by punching small holes init and that these cards might then be as-sorted and counted by mechanicalmeans according to any selected groupsof these perforations." (7) The electricaltabulating machine developed as a re-sult of these suggestions was actuallyused in connection with the 1886 mortal-ity records of the city of Baltimore, inthe Surgeon General's Office for compil-ing and tabulating Army health statisticsin 1889, and in the handling of data col-lected in the United States Census of1890.

The extension of these earliest devel-iopments to library applications wa3non-existent until the middle of the 20t1 iCentury. It is true, however, that me-chanical means for expediting librarywork have been considered for manyyears, although not in what would gv'm-erally be considered the data processingarea. The very first volume of the fore-runner of the Bulletin of the Medical Li-brary Association, the Medical Libraryand Historical Journal, dated noa, con-tained an article on the use of the " e-writing machine" in cataloging. In a di-tion, telephones have long been accept-ed as library machines, and, eicept forunusual items, hand bookbinding hasgiven way to machine-bound' boolcs. Itwas not until the late 1940'u, however,

that a few libraries began using unit rec-ord equipment for their various businessoperations such as purchasing and pay-roll.

As computers became available, a fewindustrial and research libraries beganlimited information storage and retrievalwork using these devices. Basically,these early systems closely resembledthe processing done with the usual pay-roll and personnel records formattedfiles, read serially. The development ofthe following generations of computerhardware and improved software tech-niques have led to extensive experimen-tal work searching for effective applica-tions to library operatiom. Probably themost significant advances in terms of thecomputer's application to informationstorage and retrieval have been the tre-mendous increase hi computer storageand the capability for a series of users toquery the central computer file simulta-neously and from remote locations.These developments have opened a newera for the advancement of library serv-ices.

Despite these applications and signifi-cant developmental activities in the1960's, library automation is still in itsinfancy. An exhaustive literature surveyconducted by Donald Black and EarlFarley for the publication of the firstvolume of the Annual Review of Infor-mation Scicnce and Technology (8) in-cluded veyy few articles covering as-pects of library automation systems inoperation or planned in any single li-brary. The overwhelming preponder-ance of the literature consisted of whyautomation is the coming thing, whyone should become familiar with whatmachines can do, and discussed variousapplications to one library function oranother.Automation of Library Activities

The word "srstem" as used in scientif-ic management or related to the scientif.

71

I p

ic method of problem solution desig-nates an organization of interactingparts intended to accomplish a particu-lar purpose. This rather all-encompass-ing definition can include a wide varietyof activities not the least among which isthe library. The library is certainly in-tended to accomplish a particular pur-pose, or more accurately a series of pur-poses, although these are all too oftenforgotten in the busy days of "doingthings". Certainly the library can also beproperly considered an organization ofinteracting parts.

System analysis is a method of han-dling complex problems a method sosimple and yet so powerful that it hasbeen applied successfully to problems inengineering, business, and government.If a library is truly a system and systemanalysis has proved so valuable to otherenterprises, this method of problem so-lution should be equally successful forlibrary operations. An examination of alibrary from the systems analysis view-point is desirable before considering themechanization of its activities.

Systems analysis demands the state-ment of both final objectives and of in-termediate objectives. Otherwise there isno way in which design and perform-ance can be measured an4 evaluated.Libraries may have many objectives orpurposes. There have been a number ofefforts to state these purposes in systemsterms. It can be said, as V. W. Clapphas done, . . "to provide access in bib-liographic and physical terms to the rec-ords of human communication"; (9) oras stated by Becker and Hayes . . . "teobtain, preserve and make available theprinted materials required for use by itspatrons and in so doing to store record-ed knowledge for future generations."(10) These expressions provide a back-ground against which each element orpart of the library system can be tested.

The purposes and objectives of medi-

78

cal libraries have been outlined in theopening paragraphs of this paper, par-ticularly the role of the medical libraryin medical education. Health-related ori-entation makes medical libraries dis-tinctive in that the user population isconsiderably more homogeneous than ina general library, and the emphasis thathas been placed on the library's role inmedical education requires an activerather than passive role so typical ofmost libraries. These distinctions requirecareful consideration by those engaged

analysis, design and management ofmedical libraries. In spite of thesedifferences, it is equally true that manyof the detailed functions required withinany library are basically similar.

Operationally, all library systems em-ploy four basic processes: the acquisi-tion of publications, the creation of sur-rogates for the items, the storage ofboth, and the access to one or both uponrequest. In accomplishing this work,most library practice revolves aroundthe records that are kept of the books,documents and periodicals that make upthe main store of the library. The li-brary, thus, is essentially a set of files.The basic set are the items themselves; asecond set are those providing access tothese items; and a third set are the nec-essary administrative type files. Thecreation and processing of these filesand records presents an enormous datahandling problem for the librarian. It isreadily understandable that Dr.as a librarian, would be a leader in thedevelopment of mechanized data pro-cessing equipment, but equally puzzlingthat so little use was made of it throughthe years by the libraries of the world.

The first inventory of library automa-tion activities was carried out in mid-1966 by Creative Research Services,Inc., for the Document Division of theSpecial Library Association and the Li-brary Technology Program of the Amer-

7 7

141,

ican Library Association. The report(11) tabulates the extent of mechaniza-tion and automation plans of 1130 insti-tutions of 6150 which replied to the mailquestionnaire.

The statistics resulting from this sur-vey indicate that the typical mechanizedlibrary is a univcrsity or special librarywith a collection of over 50,000 booksand more than 1,000 periodical titles. Ithas a staff of at least ten but no morethan twenty members, evenly dividedbetween professional and non-profes-sional. It has its serial control functionrunning on electrical accounting ma-chinery (EAM) and its accountingfunction running on computers. The fa-cilities of the library's host organizationare used rather than the machine equip-ment owned or leased by the library it-self. This typical library has plans tomechanize its circulation control andaccession list functions in the next oneto two years. (This description of thetypical library was derived by EugeneB. Jackson from the ;:tatistics taken fromthe survey.) (12)

Although such a typical library is onlya creation from statistics compiledthrough a survey and the data are nowalmost two years old, it does give someindication of present conditions. Perhapsmore meaningful are the statistics shownin Figure 1 for the various library func-tions. The total 638 represents the num-ber of libraries reporting some use beingmade of either electrical accounting ma-chinery (EAM) or automatic data pro-cessing (ADP) equipment. As shown bythe percentage figures, progress hasbeen made toward automating conven-tional library operations.

First EAM and now a growing num-ber of ADP systems have been adoptedfor the control of the library's manage-ment data in functions such as account-ing and in the processing activities rep-resented by acquisition, serials control,announcement lists and book catalogand catalog card production. In theareas more directly related to service tothe users, the progress toward automa-tion has been generally slower. This isthe phase of operations involving thequestion of which documents in the li-brary contain information on a specificsubject. As Adkinson and Stearns havestated, (13) in the first phase ( manage-ment data and processing activities) wehave automated our inventory controlover packages of information; in tile sec-ond phase ( the service to users), we areextending our efforts to automate intothe process of determining which pack-ages in the inventory contain answers to

Functions

Accounting 20Acquisition 12Serials Control 21Circulation Control 19Catalog Card

Production 11Book Catalog

Production 10Ar....essions Lists 15'Retrospective

Searches 12Union Lists 12Inter-Library

Communications 10Key-word-in-context

Indexes 6Current Awareness 5Number of Libraries 638

4Cr

D

1°

MEIC 2

Dim0.

e

tS'e

26 219 45

22 4814 45

7 29

15 3618 44

2413 25

2 17

20 1712 24

638 942

Figure 1 Use and plans for use of EAM and/or ADP equipment in libraries and informationcenters.

Note: Selected statistics from The Use of Data Processing Equipment by Libraries and In.formation Center.: a survey prepared fJr Document Division, SLA and Library TechnicalProgram, ALA under a grant by the Council on Library Resources; submitted by CreativeResearch Services, Inc.

79

78

particular questions. The function listedin Figure 1 as "retrospective searches"covers this latter activity and the per-centages of 12 and 24 shown in the tableare higher than one might expect. This issomewhat misleading when placed inthe restricted context of libraries; but isexplained by the reminder that the sur-vey included the information systems ofindustry where some of the more signifi-cant progress has been made in the au-tomation of retrieval operations. Thereare, of course, notable exceptions to thisgeneralization, and in the medical fieldthese exceptions exist largely because ofthe availability of the automated biblio-graphic citation file known as MED-LARS (Medical Literature Analysis andRetrieval System), established andmaintained by the National Library ofMedicine.

Medical Library Applications of Auto-mation

The automation efforts at NLM werefirst directed toward support to the pub-lication of Index Medicus. Beginning in1964, the Library placed the biblio-graphic citations for the articles selectedfor entry into Index Medicus on a mag-netic tape and used a computer togenerate a printed output. Since Augustof that year, the printing has been ac-complished with the use of the ccmput-er file and NLM's GRACE (GraphicArts Composing Equipment). This samemachine-readable file of citations nowsupports the most significant and far-reaching example of computer applica-tion to retrospective searching of medi-cal literature.

The MEDLARS file now contains ap-proximately 750,000 citations to articlesfrom about 2400 selected medical jour-nals. It can be entered through the se-lective use of the terms in a controlledvocabulary for the retrieval of the quali-fying citations. It requires trained

80

searchers to transpose a user's questioninto the proper search formulation, thepunching of a deck of cards for entryinto the computer, and then the serialsearching by the computer of the file forthe citations. This is a batch computeroperation, i.e., searches are batched ingroups of 30 to 40 and are entered to-gether for a single pass against the totalfile, print-outs are made of the retrievalsand the results forwarded to the reques-tor.

The principal operating techniques ofMEDLARS for constructing and organ-izing machine-readable records, and forthe searching and retrieving from theserecords, still require the human to assignwords as labels to represent the contentsof the documents. Then, upon searching,he must transpose a question into thesame labels and express them in logicalrelationships in order to retrieve theitem desired. The human intellect andingenuity have been put to the test toprovide effective schemes by which thismatching of terms describing the querywith those used as input descriptors canbe accomplished. The controlled vocab-ulary used in the MEDLARS system isan example of one of the more success-ful efforts to reach this goal. This vocab-ulary consists of approrimately 7000medical subject headings ( Medical Sub-ject Headings MeSH) and includes asignificant hierarchical structure to as-sist in both indexing for input and as-signing terms for retrieval.

The NLM computer capability alsosupports the publication of a series ofproducts in addition to Index Median.'Me MEDLARS file is used to producea monthly Bibliography of Medical Re-views, including articles from publica-tions such as the Annual Review seriesand Nutrition Reviews and articlestermed "review" in such journals as theNew England Journal of Medicine.These are of great value to students and

to individuals entering a new field ofmedical endeavor.

Retrospective literature searches thatare accomplished in MEDLARS are alsopublished for general distribution whentheir subject matter is judged to be ofrelatively wide interest. For example,NLM has published a literature searchtitled Computers in Medicine and Biolo-gy that contains 327 citations to articlesindexed for MEDLARS from mid-1964through December 1967. A sampling oftitles of other published searches in-cludes Tumors of Cold-Blooded Verte-brates, Neoplasm Models, Clinical Stud-ies Involving Estrogens, Epidemiologyof Suicide, and Exercise and the Elec-trocardiogram.

The NLM Current Catalog is a prod-uct of particular importance to thecountry's medical libraries. It is issuedon a biweekly basis and citations are cu-mulated quarterly. The biweekly issuescontain citations to publications cata-loged by NLM which have an imprintdate of the current or two precedingyears. Cumulations include citationsto all publications cataloged by the li-brary, regardless of the date of imprint,except for titles published before 1801and certain other special publications.In accordance with the Library's goal ofimproving and enhancing access to sig-nificant biomedical literature, the 1968Current Catalog contains, for the firsttime, separately listed citations to tech-nical reports which have been acquiredfor the NLM collection.

The National Library of Medicine iscurrently developing a complete auto-mated system as a second generation tothe present MEDLARS. The contractoris being selected to perform the systemsdesign, hardware installation, and pro-gram implementation a project that isscheduled over a five-year period. Thenew system will meet future data pro.,cessing and information system require-

81

ments for: 1) an increase in the leveland speed of MEDLARS bibliographicservices provided, including demandsearches and recurring bibliographies;2) an automated acquisitions and cata-loging system; 3) an on-line augmentedMedical Subject Headings vocabulary toaid those responsible for indexing andsearching; and 4) a drug literature pro-gram with search capabilities on chemi-cal compounds. Irt addition, the newequipment will be capable of expansionto accommodate computer processingrequired by a graphic image storage andretrieval system closely linked to theMEDLARS search capability.

Another development recently begunin computer application at NLM is theautomation of film distribution activitiesat the Library's National Medical Au-diovisual Center in Atlanta. A computer-ized system is being developed for book-ing films, preparation of shipping docu-ments, inventory and suspense control,collection of utilization data, and reportsto program sponsors. Plans call for a sys-tem which would combine source dataautomation and data processing tech-niques for maximum manipulation ofdata to produce the necessary reports,and for handling loan transactions withma;:imum speed and efficiency.

There is significant developmentalwork being conducted in other medicallibraries to apply the new computertechnology. The organization of a newmedical school library at the Universityof New Mexico has offered a unique op-portunity to develop machine methodsof bibliographic control and to use elec-tronic data processing for library func-tions. The University has been the recip-ient of a research grant from the Nation-al Library of Medicine to develop a to-tal system of library operation basedupon such automated support. The de-velopment program has included thefunctions of acquisitions, cataloging, se-

rials control, accounting, circulation andinterlibrary loan. The serials control pro-gram has shown successful use of EAMequipment for the past four years forthe activities of ordering, receiving andprocessing, claiming, producing currentlistings, and others. The circulation pro-gram includes the use of the IBM 357Data Collection System for control pur-poses. A borrower presents a pre-punched plastic identification card andthis with a prepunched book card gener-ate a transaction record. This is reportedto be an effective, accurate, fast and rel-atively simple process. Preparation ofoverdue notices is done in a matter ofminutes and management informationabout circulated items, borrowers, etc.,is made readily available.

Current work is being conducted onadapting to improved machine capabili-ties at the Washington UniversitySchool of Medicine Library, St. Louis, aleader in library automation. ( 14) In1963, this library was reporting effortsin the area of automation, specificallythe printing of acquisition lists frommagnetic tape files and the publishingof periodical holdings first from cardsand later from tapes. A year later re-ports were released on the operationalcirculation records and control systemand more recently this library has de-scribed its system for producing bookcatalogs from a single punched card in-put, manipulated by a computer underthe control of programs written for thetask. Work continues to improve theseoperations and to expand the support re-ceived from automation. For example,the acquisitions and cataloging recordsstored on magnetic tape have been ex-amined to determine frequency and av-erage length of record components andtheir alphabetization requirements in abook catalog. Data from this study arebeing used in adaptation of the library'scomputer-based cataloging system to

82

greater machine capabilities.The State Universities of New York

( SUNY) have undertaken a series ofprojects in the various areas of libraryautomation. A principal effort was theuse of a computer to support the publi-cation in 1965 of a State University Un-ion List of Serials intended to facilitatethe sharing of serial resources. The firstedition was published in the fall of 1966and contained 17,000 titles; the secondedition, dated December 1967, included25,000 titles. This publication has trig-gered a significant increase in intercam-pus exchange or interlibrary loan. Theinteruniversity transaction at the Up-state Medical Center rose from almostzero to 15-20% of the total activity at thatlibrary. SUNY has also assumed a role ofleadership in the development of abiomedical communications network,described later in this report.

The medical libraries of Columbia,Harvard and Yale collaborated on acomputerization project including theprinting of catalog cards and biblio-graphic lists. Their first product was amonthly accessions list that appeared inOctober 1963. Although this project hassince been terminated, it represented asignificant effort in cooperative libraryactivity based on computer support. TheYale University Medical Library, contin-uing with a computerization project, de-veloped a program in 1966 to producebook-form catalogs and catalog cards,using the IBM 7094-7040 Direct Cou-pled System. An acquisitions and in-process control system utilizing comput-er suppoit has also been designed byYale covering the activities from requestfor purchase to cataloging and shelving.

Information System EvaluationAre our efforts to automate accom-

plishing the intended purpose? Have wespent money for apparent sophisticationbut not really improved in the efficiency

of our job, the quality of our work or,the most important criterion, the serviceto our customer? In many cases, theonly honest answer is we don't reallyknow. The capability to evaluate the re-sults of our efforts scientifically hasproved to be a very difficult task. AsRuth Davis, Associate Director of Re-search and Development at NLM, hasstated . . . "The hopes or wishes ofthose concerned with the evaluation ofinformation systems is to apply methodsof the physical sciences to such evalua-tions." (15) She goes on to point out thefallacies of such approaches for systemswhere the human being both as anobserver and user plays such a domi-nant role. Observation itself can effectchange. The user of the information sys-tem is often considered a part of the sys-tem itself and yet represents an area ofmany unanswered questions. On theother hand, John Tukey of Princeton ex-pressed the opinion that there are, in theinformation systems area, things to dothat are obviously useful enough to needno formal detailed testing. He has con-jectured that "the real proof of the valueof change is how much it hurts if youtry to return to the old ways." (15)

There have been a limited number ofcases where major operational informa-tion systems have been subjected tolarge-scale evaluation efforts. A nationalconference on Electronic InformationHandling was held in Pittsburgh inApril of 1967. It was designed to focusattention on new ideas, research and de-velopment in the techniques of testingand evaluating electronic informationhandling systems. The report of theconference is contained in the publica-tion titled Electronic Handling of Infor-mation: Testing and Evaluation, editedby Kent, Taulbee and Belzer. (15) It isrecommended as an excellent referenceon this difficult subject. In addition, achapter has been devoted to the subject

83

of Evaluation of Information Systemsand Services and written by Alan Reesin the Annua/ Review of InformationScience and Technology, Volume 2.

(16) This covers a general discussion oftesting and evaluation as well as a re-view of the efforts in laboratory-basedtesting and evaluation, tests of operatingsystems, and research in problem areas.

One of the few major information sys-tems to "bare its soul" to the public isthe MEDLARS system. The evaluationof this system is the most elaborate andsophisticated performance study to dateof an operational system. A complete re-port of this study by Lancaster is nowavailable through the Government Print-ing Office. (17)

The study concentrated on the evalu-ation of the demand or retrospectivesearch function. Over 5,000 searches peryear are performed at NLM with addi-tional searches being run at decentral-ized centers in the United States and inEngland and Sweden. The principal ob-jectives of the test were: 1) to study thedemand search requirements of theMEDLARS users; 2) to determine howeffectively and efficiently these require-ments are being met; 3) to recognizefactors adversely affecting the system;and 4) to disclose ways improvementscan be made. The evaluation was a com-plete system evaluation since it includedall components affecting the perform-ance of the system as measured by thesatisfaction of the MEDLARS users.The study was nided and reviewed byan external board of advisers, who wereselected for their pre-eminence in thefields of information science and/or sta-tistics.

The analysis of a carefully selectedsample of 300 searches revealed that for

an "average" search somewhat overone-half of the citations in the file thatwere relevant to the request were re-trieved and that of those citations that

-"I CV" (")0 4

were retrieved between one-half andtwo-thirds were judged to be relevantby the requestor. These are consideredto be good results when compared withthe performance of other systems, al-though such results are difficult to com-pare because of the variety of methodsused to arrive at the expressions of per-formance.

A principal cause of the failure to re-trieve a relevant item was that the in-dexing was not exhaustive enough. Re-trieval of irrelevant items was due, atleast in part, to lack of specificity insearching. However, the greatest poten-tial for improvement was judged to bethe interface between the user and thesystem. Over 20% of the system failureswere traced to defective interaction inthis area. There is need for more intel-lectual effort and care in the statementof the requirements on the part of theuser. Other results highlighted include:1) need for improved vocabulary con-trol and updating; 2) 45% of input effortrequired for foreign language materialwith comparatively small acceptance ofusers; 3) bibliographic record as outputto the user is not adequate to indicatearticle content; and 4) the need for acontinuing quality control mechanismfor the future. These areas are all underpresent consideration by the manage-ment of NLM.

Networking and Cooperative Activities

There is little evidence in the litera-ture that there has been a change fromthe absence of a "total system" conceptof library automation as noted by Blackand Farley in their review, of the 1965literature. (8) They expressed the opin-ion that . . . "Were a total system con-cept ever to be implemented within a li-brary, it would certainly blur the dis-tinctions between such traditional de-partments" . . . (circulation, acquisi-

84

lion, cataloging, serials, etc. ). They con-tinued . . . "This leads one to suspectthat much of the reluctance on the partof many librarians to undertake work inautomation may stem from the fear thatautomation will force significantchanges in existing organizational struc-tures. That fear is certainly justified.However, the continued fragmentationof library operations tends to isolate theoperating divisions from their goal,which is, of course, service to a group ofusers."

It is recognized as difficult to makeradical changes in large operations butthe real advantages to be offered by au-tomation will not be realized until theentire library is recognized as a singlesystem and treated accordingly. In reali-ty, this is not even a broad enough out-look. The individual medical libraryshould be considered as a subsystem, orcomponent, of a much larger systemthe national biomedical communicationsnetwork. We believe that the local li-brary can be developed into a respon-sive biomedical communication resourceas an integrated part of an expandedcomprehensive service netwoik. Astronger national network is urgentlyneeded to serve the information needs ofall personnel in the medical sciences. Itsaim should be to permit equal and rapidaccess to all types of biomedical infor-mation.

Networks to librarians are nothingnew. Well before this word becamefashionable and Lefore the advent ofmodern computers and electronic com-munication systems, interlibrary loannetworks among libraries on a regionaland national scale were created and op-erated by members of the profession.The complex systems of national and re-gional bibliographic control in the formof union lists and catalogs, of interimsource referral services, and the largeand growing traffic in interlibrary loans

clearly identify the library system as aviable de facto network, not ordinarilyrecognized as such by its patrons. Butwith the newer tools provided by ouradvancing technology, the networktakes on a completely new dimension.

The planning, desigri and implemen-tation of the Biomedical Communica-tions Network has become a principaltask of the new research and develop-ment office at the National Library ofMedicine. This Network is the embodi-ment of NLM's plans to provide im-proved access to information within themedical community for the support ofcontinuing education, individual practi-tioners and researchers, and institutionssuch as hospitals, clinics, and researchcenters.

At least for initial planning purposes,the Network has been divided into sev-eral components by type of service to beprovided. These components are for li-brary services, specialized informationservices, specialized educational servicesand audio and audio-visual services.Supporting these four service areas willbe the data processing centers andtransmission facilities.

Planning is most advanced for the li-brary component, as might be expected.Predating this formalized program wasthe establishment of decentralized cen-ters at UCLA, Harvard, University ofMichigan, University of Alabama, Uni-versity of Colorado, and Ohio State forMEDLARS searches. Search formula-tion is accomplished at all of these cen-ters, with cards punched for computerinput. Some of the centers are actuallyrunning their own machine searches us-ing MEDLARS computer tapes provid-ed by NLM and some are sending thepunched cards for searches to be madeon NLM's equipment. Several are nowtransmitting the search requests to NLMthrough card readers and telephonecommunications links, thus, creating a

duplicate card deck at NLM for input toits computer.

An early part of the research and de-velopment program has been the estab-lishment of a Remote Information Sys-tems Center to serve both the libraryand specialized information componentsof the Network. This Center is the ter-minal facility at the Library for commu-nications links to a variety of data sys-tems and provides remote access to sev-eral computer centers and to the decen-tralized MEDLARS centers. The pur-poses of these links are to provide accessto remote data banks, support experi-mentation with the research and devel-opment of others in the information andcomputer sciences, provide a capabilityto demonstrate techniques and thestate-of-the-art in information storageand retrieval, and to assist in the datamanipulation needs of the Library. Theability to sit at a teletype in the Centerand to query a file of citations stored ona computer in California, to receive al-most instantaneous response with the ci-tation requested typed in front of youby data transmission from the computeris truly a remarkable experience. Andyet this is unsophisticated input/outputcompared with the cathode ray tubedevices now available. Such remotequery capability utilizing computer time-sharing will be a principal element of theplanned services and, hopefully, in thenot-too-distant future. The establishmentof this center represents the first tangiblemanifestation of communications net-work activity at the Library. It will serveas an operating input/output point forthe Library and specialized informationservice components of the Network.

The planning for the specialized edu-cational services places an emphasis onthe needs of continuing medical educa-tion but includes the needs of the medi-cally uninformed, particularly thosewithout the knowledge of basic and es-

85

sential family health rules. Television isthe principal medium being consideredand plans include the use of satellites toaid in transmission. This is a radical de-parture from the traditional activity of alibrary, and is an example of a need thathas been expressed for the applicationsof new technology to the medical infor-mation requirements of society.

Biomedical communications networkplanning and experimentation is also un-derway in the State Universities of NewYork (SUNY). In 1965, the SUNY ad-ministration received a proposal from atask force recommending the establish-ment of a biomedical communicationsnetwork. This network was to link theHealth Sciences Library at Buffalo, theUpstate Medical Center at Syracuse andthe Downstate Medical Center at Brook-lyn. It was proposed as a computer-cen-tered network and recommended toserve as a prototype for the University-wide library system which includes 58separate libraries. The stated purposewas to retrieve bibliographic informa-tion relating to books and journalsowned by the three libraries and by theNational Library of Medicine. Laterplans called for the addition of materialin related sciences and, finally, for auto-mation of in-house routines of the librar-ies, leading to cooperative or coordinat-ed acquisition and processing of newmaterials. SUNY accepted the proposaland designated it as a pilot study. Thenetwork is to begin operations in the fallof 1968.

Net Work development and plans areunderway in several other state librarygroups. For example, medical librarianshave recently organized the TexasCouncil of Health Science Libraries toimprove their cooperative, interlibraryactivities, and they are looking ultimate-ly to some type of network interconnec-tion. Teletype services are already beingused to support interlibrary loanactivi-

86

0

ties among five of the institutions in theNorth Texas Interuniversity Council,and in the Gulf Coast area a RegionalInformation and Communication Ex-change centering on Rice is experiment-.ing with teletype and computer servicesto participating libraries in all of thecoastal counties.

Cooperative actions among librariesof the Uniied States have too frequentlybeen limited to interlibrary loaning ofmaterials, discussions at conventions,and discussions in the literature of com-mon operating problems. Automationhas increased the awareness of the needfor improved and extended cooperationbecause of the potential advantages ofsharing work and resources through ex-change of records and linking by net-works. A new era in library cooperationwas opened in June 1967 when the di-rectors, of the three national librariesthe Library of Congress, the NationalAgricultural Library and the NationalLibrary of Medicine announced theestablishment of the National LibrariesTask Force on Automation and OtherCooperative Services. The purpose tospeed the flow of research informationto the nation's libraries and to the schol-ars and researchers who use them.

The Task Force has issued its firstprogress report which contains a list ofdesirable products and services to be de-rived from the combined efforts of thethree libr ar ie s, including bookcatalogs, machine-readable catalogingdata, demand bibliographies and coordi-nation of classification, acquisitions andreference services. The report's recom-mendations are now being used as a ba-sis for further discussions. Workinggroups have been established to attaCkeach problem or functional area andthere are signs of real progress towardunified action.

Special mention should be made ofthe progress toward the establishment of

a standard format for machine-readablecataloging ( MARC) and the distribu-tion of catalog tapes in that format bythe Library of Congress. Much of thiswork on MARC was accomplished bythe Library of Congress prior to the es-tablishment of the Task Force but nowis included in its program. Details of theproject are covered in the reports au-thored by Henriette Avram and otherslisted in the references. ( 18 ) (19)

Representatives of the three librariesare also meeting to discuss and plan fora system framework within which toplace the total effort. It is hoped thatnot only will this work lead to a forerun-ner of a truly national library networkbut that it will stimulate others in the li-brary community to increase their coop-erative efforts.Summary

The use of computers in libraries canbe summarized as supporting two mainpurposes: the handling of records gener-ated in the library's operation, and theautomatic extraction and codification ofinformation in the library's files for useby clients. The feasibility, utility, ortechnical availability of the former ap-plication is well documented. The lattertask requires considerably more researchbefore practical application will be com-mon.

The emphasis that has been placed onautomation, computers and the advancesin information science should notleave the reader with false impressionsconcerning the future of the medical li-brary. In spite of the impressive changesin the technology of information storageand retrieval, the traditional library willcontinue to serve as a vital instrument inthe educational process.

In the report prepared by the LibraryStudy Committee of the Association of

American Medical Colleges titled theHealth Sciences Library: Its Role in Ed-ucation for the Health Professions, thisposition was supported by a listing ofseveral cogent reasons. (20) Includedwere the need for the display of theprinted word through traditional meth-ods, the requirement for the housing ofcertain documents that are classics ororiginal in their presentation, and theneed for the older portion of the scien-tific record that will probably not beeconomical to transfer to some newermedia.

This is not to end, however, on a noteof status quo for the libraries. Theymust begin an active period of researchinto the most effective ways to dissemi-nate material that is of particular inter-est to the individual scientist and mate-rial that is new. The libraries must bewilling to accept an educational rolethat is much beyond where they standtoday. This should include at least theeducation of the student of the healthsciences with the form in which thescholarly record is contained and themeans by which one gains access to it,and to identify, attract, and educatenew talent for the increasingly special-ized field of information management.Finally, the health science library willhave to adopt a new philosophy in thearea of service. It must include eventu-ally the provision of linkages betweenpeople and information as well as be-tween people and documents; availabili-ty to the latest significant current events

in professional societies, conventions,meetings, hospitals and laboratories; anddevelopment of selective disseminationservices. The challenge is before the li-braries and the scientific and technicalcommunity is demanding that they ac-cept it.

REFERENCES1. Price, D. J. Science since Babylon. New Haven, Yale University Press, 1961. 149 2.2. Bush, Vannevar. As we may think. Atlantic Monthly: 101-8, July 1945.3. Dryer, B. Lifetime learning for ph:-7'cians. J Med Educ 37 (6, pt.2 ), 1962.4. Cushing, Harvey. The medical career and other papers. Boston, Little, Brown and Co.,

1940. 302 p.5. Osler, William. Selected writings of Sir William Osler. London, Oxford Univ. Press, 1951.

278 p.6. Truesdell, L. E. The development of punch card tabulation in the Bureau of the Census,

1890-1940. Washington, U. S. Bureau of the Census, 1965. 221 p.7. Garrison, F. H. John Shaw Billings: A Memoir. New York, Putman, 1915. 343 p.8. Cuadra, C. A., ed. Annual review of Information science and technology. Vol. 1. New

York, Interscience, 1966. 389 p.9. Clapp, V. W. The future of the research library. Urbana, Univ. of Illinois Press, 1964.

114 p.10. Becker, Joseph and Hayes, R. M. Information storage and retrieval: tools, elements, theo-

ries. New York, Wiley, 1963. 448 p.11. Creative Research Services, Inc. The use of data processing equipment by libraries and

information centers. Chicago, American Library Association, 1966. 160 p.12. Jackson, E. B. The use of data processing equipment by libraries and information centers-

the significant results of the SLA-LTP survey. Spec Libr 58: 317-27, 1967.13. Adkinson, B. W. and Steams, C. W. Libraries and machines, a review. Amer Docum 18:

121-4, 1967.14. Roberts, Justine. Mechanization of library procedures in the medium-sized medical library.

IV. Physical characteristics of the acquisitions - cataloging record. Bull Med Libr Assn 56:59-70, 1968.

15. Kent, Allen, Taulbee, 0. E., and Belzer, Jack. Electronic handling of information: testingand evaluation. Washington, Thompson Book Co., 1967. 313 p.

16. Cuadra, C. A., ed. Annual review of information science and technology. Vol. 2. NewYork, Interscience, 1967. 484 p.

17. Lancaster, F. W. Evaluation of the MEDLARS demand search service. Bethesda, Md., Na-tional Library of Medicine, 1968. 276 p. (In Press)

18. Avram, H. D. Implications of the library of Congress MARC project. In: Proceedings ofthe Preconference Institute of Library Automation, San Francisco, 1967. (In Press)

19. Avram, H. D. and Markenson, B. E. Library automation and Project MARC: an experi-ment in the distribution of machine-readable cataloging. In: Harrison, J. R. and Laslett,P., eds. The Brasenose conference on the automation of libraries; proceedings of the An-glo-American conference on the mechanization of library services held at Oxford, 1966.Chicago, Mansell Information Pub. Ltd., 1967. p. 97-126; discussion, p. 126-7.

20. DuVal, M. K. The health sciences library: its role in education for the health professions;report to the National Library of Medicine, January, 1967. J Med Educ 42 (8, pt. 2 ),1967.

BIBLIOGRAPHYThis list includes merely a few of the many articles and books on libraries and automa-

tion, with particular emphasis given to those items related to medical libraries and their role inmedical education. These references have been grouped under subheadings to assist those whomight wish to explore the literature more completely and in greater depth.Libraries and Information Sciences - GeneralAnthony, L. J. Library planning. In: Ashworth, Wilfred, ed. Handbook of special librarianship

and information work. 3d ed. London, Aslib, 1967. p. 233-73.Foskett, D. J. Science humanism and libraries. London, Crosby Lockwood & Son Ltd., 1964.

Kent, Xilen, ed. Textbook on mechanized information retrieval. 2d ed. New York, Interscience,1966. 371 p.

LTddider, J. C. Libraries of the future. Cambridge, Mass., M.I.T. Press, 1965. 219 p.Lipetz, B.A. Information storage and retrieval. Sci Amer 215: 224-8, passim, 1966.Overhage, C. F. Science libraries: prospects and problems. Science 155: 802-6, 1967.Salmon, S. R. Information science and automation: the newest division. ALA Bull 61: 637-42,

1967.

Taylor, S. The interfaces between librarianship and information sciehee and engineering. SpecLibr 58: 45-8, 1967.

Wooster, Harold. Books and libraries in the scientific age. Arlington, Va., Air Force Office ofScientific Research, Office of Aerospace Research, Jan 9, 1967. 23 p. (AD-645 217)

Medical Libriares GeneralArmitage, B. The present provision of medical libraries in hospitals. Postgrad Med j 42:

609-12, 1966.Ash. Lee and Bruette, V. R. Interlibrary request and loan transactions among medical libraries

of New York area. New York, Survey of Medical Library Resources of Greater New York,1966. 199

Cheshier, R. G. Medical Libraries: a changing visage. Chicago Med Sch Quart 26: 98 ,100,

1966.Cummings, M. M. The National Library of Medicine: new programs in support of medical

education, research, and practice. Clin Anesth 1: 115-23, 1966.Doe, Janet and Marshall, M. L., eds. Handbook of medical library practice. 2d ed. Chicago,

American Library Association, 1956. 601 p.Langner, M. C. The National Library of Medicine's services: initial observations on indexing,

interlibrary loan, and reference. Bull Med Libr Assn 55: 31-7 ,1967.Miller, J. C. Design for a university health sciences information center. J Med Educ 42:

404-29, 1967.Oppenheimer, G. J., ed. Regional medical library service in the pacific northwest. Seattle,

Univ. of Washington, 1967. 168 p.On, R. H. Communications problems in biomedical research: report of a study. Fed Proc 23

(5, pt.1): 1117-91, 1964.Pings, V. M. Library service for health care. Report No. 34. Detroit, Wayne State University.

School of Medicine, Library, Biomedical Information Service Center, 1967. 10 p.Services of the Boston medical library. New Eng J Med 276: 58, 1967.Libraries and Medical EducationAdams, Scott. Medical library resources and their development. J Med Educ 38: 20-7, 1963.Bloomquist, Harold. The status and need of medical school libraries in the United States. J

Med Educ 38: 145-63, 1963.Esterquest, R. T. Improving the quality of the medical school library. j Med Educ 29: 451-8,

1964.Guidelines for medical school libraries. T Med Educ 38 (1, pt.1 ), 1965.Harris, Oren. Medical libraries and medical education. T Med Edue 41: 84-90, 1966.Jeghers, H. Medical care. education and research. Philosophy and technics of self-education of

the medical student. New Eng j Med 271: 1297-301, 1964.

Long. E. C. The medical library and the medical student. Bull Med Libr. Assn 52: 568-74,1964.

Ravdin, I. S. The importance of the library in medical education. Bull Med Libr Assn 53:505-9, 1965.

Sanazaro, P. J. The unmet responsibility of academic medicine to its libraries. J Med Educ 40:65-6, 1965.

West, K. M. Role of the library in learning to learn clinical medicine. J Med Educ 39: 910-7,1964.

Wilson, M. C. and Cummings, M. M. The National Library of Medicine: relationships to med-ical education and research. J Med Edu 40: 225-32, 1965.

Libraries and Systems AnalysisBolles, S. W. The use of flow charts in the analysis of library operations. Spec Libr 58: 95-8,

1967.Corrigan, R. E. and Kaufman, R. A. Why system engineering. Palo Alto, Calif., Fearon, 1966.

72 p.Covill, C. W. Librarian plus systems analyst = teamwork? Spec Libr 58: 99-101, 1967.Markwn, B. E. System development study for the Library of Congress automation program.

Libr Quart, 36: 197-233, 1966.St. Pierre, P. L. and Chapman, E. A. Library systems analysis and design. In: Proceedings of

the Preconference Institute on Library Automation, San Francisco, 1967. (In Press)Automation and Computers CeneralFavret, A. G. Introduction to diOtal computer applications. New York, Reinhold Publishing

Corp., 1965. 246 p.

89

Gruenberger, Fred., ed. Computers and communications toward a computer utility. Engle-wood Cliffs, N. j., Prentice-Hall, 1968. 219 p.

Ka:plus, W. J., ed. On-line computing; time-shared man-computer systems. New York, Mc-Graw-Hill, 1967 336 p.

Parkhill, D. F. The challenge of the computer utility. Reading, Mass., Addison-Wesley, 1966.207 p.

Pyke, T. N. Time-shared computer systems. In: Alt, F. L. and Rubinoff, Morris. Advances incomputers. Vol, 8. New York, Academic Press, 1967. Chapt 1, p. 1-45.

Library Automation GeneralAshworth, Wilfred. A review of mechanical aids in library work. In: Ashworth, Wilfred, ed.

Handbook of special librarianship and information work. 3d ed. London, Aslib, 1967. p.524-53.

Batty, C. D. Library and the machine: selected papers and discussions from a study confer-ence held at Nottingham, April 19-22, 1966 on library applications of computers and dataprocessing equipment. Scunthorpe, Linc., Library Association, North Midlands Branch,1966. 58 p.

Becker, Joseph. Current trends in library automation. In: Proceedings of the 1966 Clinic onLibrary Applications of Data Processing. Champaign, Illini Union Book Store, 1967.

Becker, Joseph. The future of library automation and information networks (Keynote Address)In: Proceedings of the Preconference Institute on Library Automation, San Francisco, 1967.( In Press)

Christianson, E. B. Automation and libraries. Spec Libr 57: 101-2, 1966.Clapp, V. W. Closing the circuit: automation and data processing for libraries. Libr J 91:

1165-71, 1966.Covill, G. W. Automation in the library: a review of the state of the art. Spec Libr 57: 332-5,

1966.Cox, N. S. and others. The computer and the library: the role of the computer in the organiza-

tion and handling of information in libraries. Hamden, Conn., Archon, 1967. 95 p.Fasana, P. J. Determining the cost of library automation. ALA Bull 61: 656-61, 1967.Fussler, II. H. University of Chicago library automation program. In: Association of Research

Libraries. Minutes of the sixty-ninth meeting. Chicago, Univ. of Chicago Press, 1967.Geddes, Andrew. Library Automation: an essential service. ALA Bull 61: 642-6, 1967.Cenung, Harriett. Can machines teach the use of the library? Coll Res Libr 28: 25-30, 1967.Gull, C. D. The present state of library automation: a study in reluctant leadership. In: Pro-

ceedings of the 1965 Clinic on Library Applications of Data Processing. Champaign,Illini Union Bookstore, 1966.

Harrison, J. R. and Laslett, P., eds. The Brasenose conference on the automation of libraries;proceedings of the Anglo-American conference on the mechanization of library services heldat Oxford, 1966. Chicago, Mansell Information Pub. Ltd., 1967. 173 p.

Harvey, John, ed. Data processing in public and university libraries. Washington, Spartan,1966. 150 p.

Hayes, R. M. Final report on mechanized information services in the university library. PhaseI Planning. Los Angeles, Univ. of California, Institute of Library Research, Dec 15, 1967.(NSF Grant CN-503 )

Holzbaur, F. W. and Farris, E. H. Library information processing using an on-line, real-timecomputer system. Poughkeepsie, N. Y., International Business Machines Corp., 1966. 47 p.(TR 00.1548)

Institute on Library Automation, San Francisco, 1967. Library Automation; a state Of the artreview. Proceedings of the Pre-conference Institute on Library Automation held at SanFrancisco, June 22-24, 1967. Ed. by S. R. Salmon. Chicago, American Library Association,1968. (In Press).

Jackson, I. F. An approach to library automation problems. Col; Res Libr 28: 133-7, 1967.Kent, Allen, ed. Library planning for automation. Washington, Spartan, 1965, 195 p.Lilje, Pauline and Penniman, W. D. Initiation to automation. Spec Libr 57: 697-700, 1966.Lynch, M. F. Computers in the library. Nature (London) 212: 1402-4, 1966.Markuson, B. E., ed. Libraries and automation, Washington, Library of Congress, 1964, 268 p.Markuson, B. E. The Library of Congress automation program: a progress report to the stock-

holders. ALA Bull 61: 647-55, 1967.McCune, L. C. and Salmon, S. R. Bibliography of library automation. ALA Bull 61: 674-5,

678-94, 1967.

90

Overhage, C. F. Plans for project Intrex. Science 152,: 1032-7, 1966.Planning Conference on Information Transfcr Experiments, Woods Hole, Mass., Aug 2-Sept 3,

1965. INTREX. Cambridge, Mass., M.I.T. Press, 1965. 276 p.Salmon, S. R. Automation of library procedures at Washington University. Missouri Libr Assn

Quart 27: 11-4, 1966.Seminar on Automation in British Columbia University Libraries, 3d, Vancouver, 1966. Re-

port. Vancouver, Univ. of British Columbia, 1967. 8 p.Shera, J. H. Libraries against machines. Science 156: 746-50, 1907.Stewart, Bruce. Data processing in an academic library. Wilson Libr Bull 41: 388-95, 1966.Stuart-Stubbs, Basil. Conference on computers in Canadian libraries, Universite Laval, Que-

bec, March 21-22, 1966. Report. Vancouver, Univ. of British Columbia Library, 1966. 13 p.U. S. LiSrary of Congress. A preliminary report on the MARC (Machine-Readable Catalog)

Pilot Project. Washington, 1966. 101 p.Warheit, I. A. Current developments in library mechanization. Spec Libr 58: 420-6, 1967.Working conference on computer applications for public libraries; proceedings. As lib Proc 18:

238-76, 1966.Medical Library Automation GeneralBrodman, Estelle. Machines in medical libraries. Med Biol Illur, 16: 238-42, 1966.Cain, A. M. Steps toward a computer-based library network: a survey of three medical librar-

ies.-tull Med Lan. Assn 55: 279-89, 1967.Divett, R. T. There's nothing like pushing a button. Bull Med Libr Assn 55: 324-8, 1967.

Pings, V. M. Obstacles in automating the health science library. Hospitals 41: 55-8, 166, July

1, 1967.Pizer, I. H. Automation in the library. Hos), Progr 47: 65-72, 1966.Library Automation AcquisitionsDunlap, Connie. Automated acquisitions procedures at the University of Michigan library.

Libr Resources Tech Serv 11: 192-202, 1967.Dunlap, Connie. Automated acquisitions systems. In: Proceedings of the Preconference Insti-

tute on Library Automation, San Francisco, 1967. (In Press)Miller, E. F. and others. Automated book ordering and receiving. spec Libr 57: 96-100, 1966.

Weis:, I. T. and Wiggins, E. V. Computer-aided centralized cataloging at the National Library

of Medicine. Libr Resources Tech Serv 11: 83-96, 1967.Library Automation CatalogingCartwright, Kelley. Automated production of book catalogs. In: Proceedings of the Preconfer-

ence Institute of Library Automation, San Francisco, 1967. (In press)

Cronin, T. W. Centralized cataloging at the national and international level. Libr Resources

Tech Ser 11: 27-49, 1967.DeGennaro, R. Strategy for the conversion of research library catalog to machine readable

form. Coll Res Libr 28: 253-7, 1967.Kilgour, F. G. Library catalogue production on small computers. Amer Docum 17: 124-31,

1966.Kilgour, F. G. Mechanization of cataloguing procedure. Bull Med Libr Assn 53: 142-62, 1965.

Lane, D. 0. Automatic catalog card production. Libr Resources Tech Serv 10: 383-6, 1966.

Roberts, Justine. Mechanization of library procedures in the medium-sized medical library. V.Alphabetization of the book catalog. Bull Med Libr Assn 56: 71-9, 1968.

Simonton, Wesley. Automation of cataloging procedures. In: Proceedings of the Preconference

Institute on Library Automation, San Francisco, 1967. (In Press)

Library Automation CirculationPizer, I. H. Mechanized circulation system. Coll Res Libr 27: 5-12, 1966.

Stockton, P. A. IBM 357 circulation procedure. Coll Res Libr 28: 35-40, 1967.

Library Automation Citation RetrievalAdams, Scott. MEDLARS: performance, problems, possibilities. Bull Med Libr Assn 53:

139-151, 1965.Cox, N. S. and Grose, M. W. Organization and handling of bibliographic records by computer.

Newcastle upon Tyne, Oriel, 1967.. 187 p.Darling, Louise. The MEDLARS regional center at UCLA: a progress report. Bull Med Libr

Assn 54: 311-5, 1966.Hogan, Rose. An evaluation of MEDLARS output: demand and recurring bibliographies. Bull

Med Libr Assn 54: 231-4, 1966.

91

04_ ,:t.

7.7.51V.?:1Veltsmirager.

Kilgour, F. G. Symbol-manipulative programming for bibliographic data processing on smallcomputers. Coll Res Libr 27: 95-8, 1966.

Mathews, W. D. The TIP retrieval system at M.I.T. In: Schechter, George, ed. Informationretrievala critical review. Washington, Thompson Book Co., 1967. p. 95-108.

The MEDLARS story at the NaConal Library of Medicine. Washington, U. S. Dept. ofHealth, Education, and Welfare, 1983. 74 p.

Rogers, F. B. MEDLARS operating experience at the University of Colorado. Bull Med LibrAssn 54: 1-10, 1966.

Rogers, F. B. MEDLARS operating experience: Addendum. Bull Med Libr Assn 54: 316-20,1966.

Truelson, S. P. What the Index Medicus indexes, and why. Bull Med Libr Assn 54: 329-36,1968.

Library Automation SerialsCurran, A. T. Mechanization of the serial records for the moving and merging of the Boston

medical and Harvard medical serials. Libr Resources Tech Serv 10: 362-73, 1986.Lebowitz, A. I. The AEC library serial' record: a study in library mechanization. Spec Libr 58:

154-9, 1967.Payne, L. M. and others. Mechanization in a new medical school library: serials and circula-

tion. Bull Med Libr Assn 54: 337-50, 1968.Stewart, Bruce. Automated serials system. In: Proceedings of the Preconference Institute on

Library Automation, San Francisco, 1967. (In Press)Wilkinson, W. A. A system for machine-assisted serials control. Spec Libr 58: 149-53, 1937.National Plans and NetworksBrown, G. W., Miller, J. G., and Keenan, T. A. EDUNET; report of the summer study on

information networks conducted by the Interuniversity Communications Council. New York,Wiley, 1967. 440 p.

Cummings, M. M. The biomedical communications problem. In: DeReuck, Anthony andKnight, Julie, eds. Ciba foundation symposium on communication in science: documentationand automation. London. J. & A. Churchill, Ltd., 1967. 110-22 p.

Mohrhardt, F. E. and Oliveri, B. L. A national network of biological-agricultural libraries. CollRes Libr 28: 9-16, 1967.

Pizer, I. H. The State University of New York computerized biomedical information resource.In: Harrison, J. R. and Laslett, P. eds. Brasenose conference on the automation of libraries;proceediings of the Anglo-American conference on the mechanization of library services heldat Oxford, 1966. Chicago, Mansell Information Pub. Ltd., 1967. p. 151-62; discussion, p.162-4.

Skipper, J. E. Interinstitutional cooperation in library services. Liberal Educ 54: 65-72,1968.

Weinstock, Melvin. Network concepts in scientific and technical libraries. Special Libr 58:328-34, 1967.

92

-

SUMMARYCOMPUTERS IN MEDICAL LIBRARIES°

Panel: Martin M. Cummings, M.D., Chairman, Irwin Pizer, Robert T. Divett, Leonard Eddy.

The panel on Computers in MedicalLibraries was pleasantly surprised tofind an audience of approximately 30participants, and the word "partici-pants" should be emphasized becausemore than half of the alloted time wasspent in active dialogue between partici-pants and panel members. In an attemptto meet the charge, a statement of gen-eral purpose has been developed.

The importance of the medical libraryas a learning resource for education isreflected in many ways. Local, state andfederal institutions and organizationsare deploying more funds and manpow-er in an effort to restore this instrumentof education and learning to a tool of ef-fectiveness and efficiency geared tomatch the expansion of the total educa-tional needs of our nation. Libraries areseizing this opportunity to use the com-puter and related new technologies tosolve some of the existing informationproblems. There is still need, however,for refinement and further discrimina-tion in application of computers to thisenormous set of communications prob-lems. The most encouraging evidencethat success is close at hand comes fromconferences such as this one in whicheducators, administrators and librariansare sharing views, experiences and rela-tionships critically, candidly and cooper-atively. To share resources on a nationalbasis, plans for developing local, region-al and national library networks seem to

*Presented by Martin M. Cummings, M.D., Panelconference.

93

be advancing at a rapid pace. The needto consider standards for compatibilityamong computer systems has been rec-ognized and much work is underway tobridge program and hardware dispari-ties in an effort to effect economy andefficiency of operations.

It has been estimated that more than90% of all health personnel that have everdedicated their lives to research arealive and working in laboratories andclinics today. As a result, the immensegrowth of knowledge and printed mattersince World War II has caused a greatconcern among those who need to findand use this information. In 1963, it wasestimated that the growth rate of publi-cations over the previous 20 years hadbeen about 3.1% per year, doubling every22 years. More recent estimates reflect adoubling every 11 years. It was furtherestimated that information in the world'slibraries is growing at a rate of 2 x 10°

bits per second (a bit in this contextbeing approximately equivalent to halfof one letter or one numeral). Thisgrowth has created severe strains thatare beginning to break down the effi-ciency of all methods of informationhandling that heretofore have been rea-sonably adequate. As a result, a newmethodology of information handling isdeveloping. This methodology is basedupon the ability of computers to per-form some of the routine functions of in-formation management and processing

Chairman, at the closing plenary session of the

previously handled by man alone. Thepredecessor of the computer was thepunch-card accounting machine devel-oped by Hollorith and Billings, and theuse of the punch-card equipment in li-braries goes back to the early 1930's.Since that time, a few libraries havebeen using this methodology for some oftheir routine library procedures. Amongthe first libraries to use punch-cards wasthe University of Texas where RalphParker developed a punch-card circula-tion system, since adopted by many oth-er places.

There have been two schools ofthought concerning the use of comput-ers in information handling. The librari-an, for the most part, has looked at thecomputer simply as a means of accom-plishing the traditional tasks of libra-rianship. The documentalist, on the oth-er hand, has looked at the computer as apotential means of storing, retrievingand manipulating information of alltypes. Vannevar Bush in 1945 publishedthe provocative paper which has led tothe most revolutionized thinking ofthose who are concerned with informa-tion management. From his report hasgerminated the concept of fact retrievalas opposed to citation retrieval, whichhas always been the librarian's ap-proach. The ability to retrieve citationsis a necessary prelude, however, to anyeffective machine retrieval of actualfact. Except in a very few cases, thetechnology of information science hasnot yet developed to the point wherethere can be fact retrieval by machine.The National Library of Medicine hashad some success with its medical litera-hire analysis and retrieval system, butits principal contribution has been to in-troduce computer technology on a fairlyextensive scale within the library set-ting. It is now planning the develop-ment of a second MEDLARS systemwhich will allow the use of time-shared

94

computers in retrieving not only refer-ences or citations, but abstracts and se-lected portions of total text as well.

Serious investigation of the use ofcomputers for all library operations be-gan with the development of the newcampus of the University of Illinois inChicago. Shortly thereafter, one of theleaders of the team at Illinois acceptedthe position as Director of Library andInformation Services at the new FloridaAtlantic University in Boca Raton. Dr.Heiliger put into effect at Florida Atlan-tic the plans previously developed at theChicago undergraduate campus at theUniversity of Illinois. I think we shouldpoint out that the Florida Atlantic pro-ject was not an unqualified success. Infact, some would say it was an abysmalfailure and the panel believes that refer-ence to failures is important to recalland that we should not overemphasizeour successes. Shortly after the Illinoisreport Dr. Estelle Brod.man and Mr. Ir-win Pizer began a project aimed at totalsystem library operations at WashingtonUniversity School of Medicine in St.Louis.

Selected library operations in thisproject have been developed successful-ly using the batch mode of computer op-erations. About the same time, Fred Kil-gour, Librarian at the Yale Medical Li-brary, conceived of a project that wouldplace the holdings of his great Library,those of the Harvard Medical School Li-brary, and those of Columbia UniversityMedical Library onto a central comput-er storage device connected to each li-brary by leased lines and typewriter ter-minals using time-sharing principles. Al-though Kilgour had apparently workedcarefully and meticulously to developthe software for this program, this effortfailed for an entirely different reason.Harvard withdrew from this project,presumably for administrative or organi-zational reasons, so there is a second evi-

0 0

&nee of failure, but in this case appar-ently not caused by a lack of technicalplanning.

One cannot attempt to review devel-opments in this field without makingreference to the exciting work which hasbeen taking place at the MassachusettsInstitute of Technology during the pastdecade and, in the interest of time, willonly refer to the accomplishments ofproject MAC and its related and subsid-iary project TIP which has demonstrat-ed the ability of users to work on-linewith massive data files generated, infact, by the users themselves. The con-cepts emerging from this experimenthave been seized upon by other groupsin the evolution of subsequent on-line li-brary and data-based information devel-opments and perhaps the most signifi-cant of these is the second M.I.T. proj-ect, INTREX, which will attempt, be-tween now and 1975, to develop a factretrieval library-based information sys-tem within the field of interest of engi-neering and other hard sciences.

Returning to medicine, in 1963 theUniversity of New Mexico began to es-tablish their new school. The Universityof New Mexico School of Medicine is arelatively young school, graduating itsfirst class in June, 1968. The School ofMedicine Library was faced with thetask of building an adequate library in ashort time. It was early decided to uti-lize machine methods to the fullest ex-tent possible. The first few months werespent in designing a punch-card systemand in putting records in machinable.form. It was found that no total systemwas available for use in the Library. Itwas necessary to borrow, modify, andredesign several programs to arrive atusable solutions to problems encoun-tered. This punch-card system has beengradually phased out in the pastmonths, and is being replaced by batch

95

processing on an IBM 360 Model 40Computer.

New Mexico is now doing research todesign a total library on-line systemone which can, with only slight modifi-cation, be used in other libraries. Onerequirement is that these programs bewritten as nearly as possible in a univer-sal language. An IBM 357 Data Collec-tion System is now being utilized for cir-culation procedures, and a daily print-out is obtained, much as is done in manylibraries across the country.

Planning and implementing of on-lineprocedures is well under way. Two IBM226 CRT Terminals will be installed inthe Library in June of this year, to beconnected to the IBM 360 Computer.These terminals will be utilized in ac-quisitions and cataloging processes, aswell as for serials control and circulationprocedures.

Acquisitions and cataloging' programswill be closely interrelated when theCRT Terminals are in operation. On re-ceipt of a book request, the acquisitionslibrarian will place this record on thecomputer. From this time until final ca-taloging, the original record will be re-called and supplemented or corrected asnecessary on the CRT Terminal.

The serials control program is similarto acquisitions procedures. A descriptivecataleg entry will be made for each se-rial title. Receipt of a serial will be re-corded on the CRT much as is nowdone in a Kardex file. Claiming proce-dures are more difficult. The vagaries ofpublication and distribution of serialsmake necessary the development of aman-machine interface to effectively ac-complish this necessary claiming opera-tion. There is an anticipation period foreach serial, a time period within which aparticular issue is expected to arrive.This period might be 30, 60, or 90 days,for instance. Should the item not be re-ceived during the ant'cipation period,

this fact will be brought to the attentionof the serials librarian. She will decidewhether to claim any particular serial,and a completed claim will be producedwhen required.

The primar y impact on the user of theLibrary will be that, instead of going toa card catalog, he will go to a cathoderay tube terminal to begin a literaturesearch. The computer will be pro-grammed to ask questions of the user.These questions and his answers will,hopefully, duplicate the thought proc-esses which take place in use of a tradi-tional catalog. The user should be ableto locate needed information in less thana minute. Appropriate citations will thenbe displayed on the terminal or printedout on an attached printing unit as de-sired. One additional search method willbe possible with the software being de-veloped a TWX search. Any teletypestation should be able to dial into thecomputer system and do a literaturesearch to any depth.

For all practical purposes, the com-puter system and library being devel-oped by the University of New MexicoSchool of Medicine is not a local system,but rather is a statewide or even region-al information center.

This model was discussed in somedepth by the panel with a view of deter-mining whether a system developed atone new medical school might be trans-lated or transferred to other new medi-cal schools.

The panel then spent much time re-viewing the interesting developments ona state-wide medical library networkwhich is being developed by the StateUniversity of New York.

The State University of New YorkBiomedical Communications Networkbegan in 1965. The Network now linksState University Medical Center Librar-ies at Buffalo, Syracuse, Brooklyn andStony Brook. In addition, the University

of Rochester Medical Center and Alba-ny Medical Center Libraries have joinedthe Network. The collections of these li-braries total more than 600,000 volumesand more than 9,000 serials are currentlyreceived.

An IBM 360 Model 40 Computer withIBM 2740 Communication Terminalsare utilized in the Network. Exchange ofinterlibrary loan data, reference requestsand other administrative messages con-cerned with the libraries and the Net-work are facilitated as secondary appli-cations of the Network.

The Network is primarily oriented tothe user with goals and functions clearlystated as follows:1. Production of a computerized catalog

of monographs in the three SUNYmedical libraries, published from1962 onward.

2. Production of union lists of serialsand books with periodic updating(this may include the medical librar-ies, the entire 58 campus SUNY Li-brary System, or a combination of li-braries within the system).

3. Search of the combined catalogs forall libraries for monographs bearingimprint dates of 1962+. Eventually10 years of book data will be storedon-line. The initial data base consistsof approximately 20,000 records.These can be searched by author, ti-tle, subject, and other special fields ifdesired.

4. Search of 5 years of MEDLARS data,to be obtained from the National Li-brary of Medicine. This amounts toapproximately 1,000,000 records, alsostored on-line. These will be able tobe searched by author (name) orsubject, with the usual limitingparameters of time, language, sex,date, etc.5. Search of the SUN? Union List of

Serials data consisting of over 20,000current and non-current periodical titles

held in the 60 libraries of the State Uni-versity, plus Syracuse University, Hamil-ton College, Colgate University, andseveral smaller schools and special li-

braries.6. Preparation of lists of currently re-

ceived titles for each campus, or the en-tire system, of books arranged by au-thor or subject, and of journals ar-ranged by title or subject, or any desiredcombination. The information dissemi-nation aspect of this phase is of greatimportance to all faculty, research staff,and student users.

7. Education and training of librarypersonnel and orientation and trainingfor users.

8. Current awareness services Selec-

tive Dissemination of Information(SDI). Individual faculty members orresearch groups could be notified ofnew articles and books in their special-ized fields of interest as each newMEDLARS tape is received and as cata-log information is added to the masterfiles.

9. Key Word in Context (KWIC) in-dexing of appropriate local researchdata, and, if desired, of journals not cov-ered by the major indexing services.

10. Recurring bibliographies, e.g.,monthly for a specific research group orproject.

11. Research and development oftechniques for information storage andretrieval (see below ).

The major elements of the data basewill be augmented by the circulationrecords for the Upstate Medical CenterLibrary and the other libraries as theydevelop automated circulation systems.

The user will query the system him-self, not through an intermediary, in hisown terms, which will be translated, ifnecessary, internally. Within two min-utes of the completion of his querystatement he will begin receiving the re-sponses.

97

NLM Shared Cataloging ProgramIn 1966/67 the National Library of

Medicine negotiated agreements to be-

gin expanding the scope of the CurrentCatalog. As stated in the 1967 fiscal yearAnnual Report of NLM:

Negotiations were completed for amanually shared cataloging projectwhich will result in the Francis A.

Countway Library's acquisitions beingincluded in the CURRENT CATALOG.In FY 1968, the Francis A. CountwayLibrary will begin submitting masterforms of their catalog citations for inclu-

sion in the CURRENT CATALOG.Those which are added to the catalogwill reflect the NLM classification andMESH subject headings, as well as theCountway Library symbol, MBCo.In addition,

SUNY/NLM Shared Cataloging Proj-ect arrangements were completed withthe SUNY Upstate Medical Center, Syr-

acuse, New York, to:A. Convert NLM current cataloging

tapes to IBM 360/40 and to refor-mat the NLM data according to thecurrent tape tormats to projectMARC of the Library of Congress.

B. Provide programs for searching ofthe resultant tapes using IBM's GISdocument handling module.

C. Experiment with batch searches ofthe resultant tapes.

D. Share input in machine catalogingdata.

E. Cooperate in designing a shared ca-taloging system.

Beginning in April, 1968 both Count-

way and SUNY items will appear in theCurrent Catalog, and the contract be-

tween SUNY and NLM has been re-newed in order to compile and search amaster data base consisting of NLM,Countway, and SUNY machine readablerecords.

A total of 32 full-time persons are cur-rently engaged in the implementation of

"IM:nraiew,..W5errnierenren*.m,e.-rnsuynAATI.Vrwmln,..rvezn,Ifotor.,,,,..,1,1.2.1.11-,,,,,,...ya,m,

the Network, in addition to the existingstaffs of the component libraries whonumber 92 and who may be thought ofas resource persons for the Network pro-gram.

Par all eling this development ofmechanization of traditional libraryprocedures has been a significantchange in the concept of librarianship,especially medical librarianship. The li-brary function has been defined as theconscious interruption of the totalstream or flow of recorded ideas and in-formation, the selective drawing off ofmanageable amounts for storage, laterretrieval and distribution to individualsor groups in whatever media forms orformats might be required and are ap-propriate to satisfy known or anticipatedneeds. In other words, libraries will beIncreasingly concerned with manage-ment of information in whatever formatit is in or the media used to disseminateit.

In response to the question "Whatwill libraries look like in the future?",Stone has stated the library of the futureis not wisely conceived as a single placeat all, but rather as a far-flung networkcomposed of units of various sizes andshapes, each of which may perform sim-ilar as well as different functions, but allof which will be linked together elec-tronically or mechanically. In the future,it will probably be less necessary tohave all of the pieces of a library pro-gram in one place so long as the pro-gram parts can be linked together. Inother words the concept has developedthat the library should be a place fromuser, rather than a place where the userwhich information is disseminated to themust come to acquire the information.

The trend seems to be developing,particularly in newer medical schools,toward the centralization of administra-tion of all agencies concerned with themanagement and communication of in-

98

formation. Among those places wherecomputers, medical illustration activi-ties, educational media developmentand library science operations are con-solidated under one administrative headare the University of New MexicoSchool of Medicine, the University ofTexas Medical Schools in Dallas andSan Antonio, the California College ofMedicine of the University of California,the University of Arizona School ofMedicine and more recently, the Univer-sity of Alabama Health Center. In thosecases where librarians are qualified toserve as information scientists, the li-brarians have been placed in charge ofthis administrative unit. In other places,men from science or medicine who havebecome interested in the modalities ofinformation transfer are assuming thisimportant role.

The consolidated operation lends it-self toward the development of the tech-niques of information handling that arenow on the horizon, but not sufficientlydeveloped for implementation for five orten years in the future. Among these arethe methodologies of fact retrieval ear-lier referred to. Such techniques mayuse microfilm, microform, video tape,motion pictures, facsimile transmissionand other media. If there is no centralmanagement, implementation and devel-opment of medical information systemswill be hampered and, indeed, might bechaotic in terms of potential linkages.

At the national level, the National Li-brary of Medicine has underway a mas-sive systems design for biomedical com-munications which involves several ma-jor information components. They are(1) an automated document handling li-brary network, (2) the development ofa network involving the National Medi-cal Audiovisual Center and its concernwith other learning resources, (3) sys-tems management through monitoring,control feedback and application of con-

trol system procedures, (4) evaluationof systems operations so as to extractsystems parameters, (5) systems inade-quacies and component coupling, and(8) the support of education throughautomated information and communica-tion networks, including the possible useof satellites for public as well as formedical education. All of these effortsare related to establishing compatibilityamong the three national libraries, aswell as with specialized informationcenters and other emerging national in-formation programs. Medicine presentlyis in the forefront of these efforts and in-tends to share fully its experience withother disciplines of science and technol-ogy. The methods developed by NLMfor decentralizing its computer-based in-formation programs and, particularly, itsextension into all levels of support formedical research, education and prac-tice has led the President to include it asa prototype for the developing "knowl-edge network." In this context, medicalschools, professional societies and otherhealth related organizations must partic-ipate by providing evaluated substantiveinput of high quality and pertinence tothe major health problems of the nation.

It was clearly demonstrated at thepanel meeting that medical librarieshave been in the forefront in the investi-gation and installation of modern ma-chine methods to cope with the risingflood of materials being published. Li-brarians have felt the need to preparethemselves to more fully utilize comput-ers in documentation and control of do-cumentation. It was suggested that thelibrarian take advantage of computertraining on his own campus, becausefrom the expressions at this meeting itbecame clear that computer center per-sonnel are excellent sources of informa-

tion and support. In addition, attend-ance at computer courses, workshopsand other outside advanced trainingprograms should be encouraged when-ever possible. Libraries should seek toemploy systems and programming _per-sonnel as early as possible in any com-puter planning development. Sharing ofinformation, ideas and programs hasgiven impetus to the development ofseveral important cooperative ventures.

Although some librarians apparentlystill believe that students, faculty andother library users might resist the useof computer terminals in their search tolocate information, it was pointed outthat these students are now being earlyintroduced into instrumentation andcomputer technology in their medicalschool careers. Therefore, the libraryshould be considered as another re-source to acquaint the student with dif-ferent types of equipment with whichhe may be interacting in his futurework.

The final recommendation of the pan-el was that continued federal privateinterest and support for improved medi-cal library service should be sustainedand, wherever possible, expanded. How-ever, careful review of on-going systemsdevelopment in libraries should be madeto avoid duplication of effort and thusavoid unnecessary expenditures.

The panel was impressed that costsharing and program sharing is now tak-ing place at all levels in the tradition oflibrary cooperation which has developedover the past century. Although soft-ware and hardware are becoming easierto acquire, more attention should be giv-en to the development of the liveware",the people, needed for the ,effective useof these systems.

'1,:!"..":".rf:T7Tr:vrromnr-resew.

A PROGRESS REPORT OF COMPUTER ASSISTEDINSTRUCTION AT THE UNIVERSITY OF

OKLAHOMA MEDICAL CENTER*

Panel: William G. Harless, M. S., Chairman, Thomas N. Lynn, M.D., Robert C. Duncan, Ph.D.,Eugene D. Jacobson, M.D., James A. Cutter, M.D., Arthur W. Nunnery, M.D.

In the fall of 1966 the University ofOklahoma Medical Center ComputingFacility initiated implementation ofcomputer assisted instruction ( CAI) asan educational capability for the faculty.The first year was used primarily for theindoctrination of a staff and the explora-tion of the pedagogical possibilities ofthe CAI technique. These activities havebeen previously reported.

The first step in a new installationmust be to define computer aided in-struction from a philosophical point ofview. When this is done two very prag-matic problems become apparent,WHERE and HOW it fits into the aca-demic goals and activities of the Univer-sity. These questions must be answeredbefore any significant implementationtakes place. This answer requires a tax-onomy of medical education. For con-venience we have divided it into fourmajor areas:

1. Graduate education in the medicalsciences

2. In-service training in health relatedfields

3. Undergraduate medical educationand house staff training

4. Continuing medical educationOur plans include activity in all of

these areas. The following discussionwill concern the various activities ofCAI in several disciplines and reflectsthe interest in the four areas mentionedabove. Only information concerning

material presently developed and opera-tional is discussed.

The use of CAI in graduate educationin the medical sciences had its begin-ning in a course called "Medical Back-grounds." The course was developedthree years ago to accommodate the in-creasing enrollment in the graduate divi-sion of the Department of PreventiveMedicine and Public Health. Studentsentering the disciplines of environmen-tal health, health administration andbiostatistics needed additional back-ground in order to converse with physi-cians and other persons in the healthprofessions. The course, "Medical Back-grounds," was therefore designed prima-rily to facilitate communications be-tween these students and others in thehealth profession. The course evolvedinto a two-hour didactic course lastingone semester.

The IBM 1401 Data Processing Sys-tem was restructured to permit imple-mentation of computer assisted instruc-tion utilizing four 1050 typewriter termi-nals. It was decided that "Medical Back-grounds" was appropriate to augmentthe instruction by means of this educa-tional technique. Ten lessons were de-veloped 'utilizing the drill approach toreinforce the material covered by tradi-tional lecture. Participation in the com-puter assisted instruction portion of thecourse was voluntary, and the studentreactions were mixed. Some students

Presented and demonstrated by faculty members involved in computer assisted instruction.

100

found it fascinating; others resented it.Many mechanical failures such as thepaper getting jammed in the typewriterterminal and programming systems fail-ures were constant sources of frustrationto the students.

The following year the IBM 1500 In-structional System was available for thereinforcement drill sessions. This elimi-nated many of the machine problems ofthe first year and proved to be a muchmore effective presentation. The greaterflexibility in presentation afforded bythe cathode-ray tube displays made thewhole effort a much more palatable onefor the students.

There was an effort at evaluation ofthe results during the second year. Thefollowing observations were made con-cerning grade distributions and their re-lationship to the number of CAI lessonstaken:

Grade

A

Number of Average CAIStudents Lessons Taken

10 8.824 6.4510 5.5

1 2.0

Perhaps the most plausible explana-tion for this supportive data is that theA students were the more highly moti-vated students who can be expected toavail themselves of any educationalmaterials available.

The following is a list of the drill ses-sions available to the students in Medi-cal Backgrounds: 1. Medical Terminolo-gy; 2. Cellular Structure; 3. Dermatolo-gy; 4. Muscular-Skeletal System; 5.Hematology; 6. Respiratory System; 7.Endocrine System; 8. CardiovascularSystem; 9. Gastrointestinal Tract; 10.

Urogenital System.There is a lesson presently under de-

velopment concerned with the spinal

101

cord. It is a tutorial approach to theteaching of the basic components andfunctions of the spinal cord, and is de-signed to replace a lecture as opposed tothe previous ten which serve as supple-mentary drill sessions for material pre-sented by lecture.

One of the more promising areas ofcomputer assisted instruction seems toexist when a student is allowed to pro-ject his existing knowledge of a subjectinto a simulated situation. This area isrelatively unexplored for many reasons.Currently it is technically difficult to in-terface effectively the problem solvingand tutorial techniques in the CAI envi-ronment. Another difficulty is the inabil-ity of the author to adequately definethe pedagogy involved in such a tech-nique which results in dangerously ne-bulous goals for the session. One furtherimportant problem that inhibits theprogress of such development is the to-tal involvement required of the profes-sor during the development of the ses-sion.

In an attempt to transcend theseproblems and provide an effective learn-ing tool, the Department of Biostatisticsand Epidemiology of the School ofHealth is responsible for preparing a les-son in sampling theory involving a dy-namic model of a population. The lessonrequires an interface between the Course-writer langliage for a tutorial modeand the MAT language for a problemsolving mode. The simulated situation isdeveloped for the student in the Course-writer mode. He is allowed to pick whathe considers to be the appropriate ana-lytical methods needed to solve the sam-pling problem presented. Once this isaccomplished he is directed to switch tothe MAT mode. Utilizing Monte Carlomethods in this problem solving mode, asample is generated. The student's pro-cedural choice in phase one is used, andthe appropriate subroutines are called

-.* ^

by the student to analyze the generateddata. When the analysis is complete, thestudent switches back to the Course-writer mode Acre structured interactionoccurs between the student and thecomputer concerning the appropriate-ness of the student's analysis andother aspects of the problem. This isthe critique phase of the student's prob-lem solving experience and the tutorialmode will again be in effect. If the stu-dent choose the appropriate methods, hewill be reinforced. If his chosen metho-dology was inappropriate, he will betold so and a proper method suggested.The problem may then be extended or anew problem introduced and a branch isexecuted hack to phase one for anotherchoice of methodology by the student.Tbe plans arc for various data to be cop-ied by the student and brought back forclass discussion. This allows for manydifferent sampling situations to be pre-sented and facilitates class discussion.

It is important that the students ap-preciate the dynamic nature of the situa-tion. They must not feel that the situa-tion is so structured that there is a singlesolution to any problem presented. Thissystem gives him tbe feeling of problemsolving in real time: a live situationrather than guessing with the machineor the professor concerning the answerto a structured problem.

These techniques may be utilized inother areas of medicine such as diagnos-tic procedures, patient management,computation of water and electrolytebalances, drug therapy and the compli-cations of misuse of drugs, and possiblyin epidemiology concerning the investi-gation of contagious disease models.

It is very difficult indeed to measurethe efficiency of such an educational ap-proach and, as has been stated earlier,the total involvement required of a pro-fessor is an inhibiting factor to this ap-proach. But the knowledge that one is

teaching things that are difficult toteach under the current system is per-haps justification enough for the effort.More work needs to be done concerningtechnique development before appropri-ate evaluations and traditional efficiencymeasures can be made.

The Department of Pediatrics haselected to use the tutorial approach inprogramming lessons concernig theprinciples of feeding in infant nuvitionfor junior and senior medical stucl,mts.Under the current curriculum the 6irdyear students rotate in groups of 12 or13 for five weeks on the pediatrics serv-ice. It is during this period of time thatthey are exposed to the computer assist-ed instruction material covering infantnutrition. The fourth year medical stu-dents rotate in groups of two for oneweek in the newborn service. These stu-dents are responsible for talking withthe mothers of the newborn with refer-ence to the feeding requirements, selec-tion of the formula, and explaining theappropriate preparation procedures.The computer assisted instruction les-sons on formula feeding and infant nu-trition are used as orientation sessionsfor these students. The decision hasbeen made to pursue infant nutrition ingreater depth utilizing CAI techniques.Information concerning molecular nutri-tion taught by the Department of Bio-chemistry will soon be available utiliz-ing the computer assisted instruction fa-cility. Ultimately, it is envisioned thatby the time a freshman medical classreaches the third year, all of the infor-mation concerning these matters relatedto infant nutrition in the basic scienceswill be available to the students forreference or as refresher courses. Hope-fully this will help to bridge a gap be-tween the basic science education andthe clinical experience.

Generally, at this point, the accept-ance by the students involved is favora-

102

10 I

ble. Previously a two-hour-a-week tuto-rial session was requircd in infant nutri-tion by the Pediatrics instructors withthe fourth year students. This has beencut to about ten minutes a week.

One of several unique features addedto the course in medical physiology forthe academic year 1968 is the use ofcomputer, assisted instruction to teachtwo lessons. The lessons concern "theaction potential" and "pancreatic secre-tion." The Physiology Department hasmoved cautiously into the area, carefullybuilding in evaluation procedures foracadcmic effectiveness. For. the materialconcerning the action potential, one halfof the class is randomly selected to takethe lesson on the computer; the otherhalf of the class takes the lesson by tra-ditional lecture. Some weeks later forthe material on pancreatic secretion thetwo halves of the class are reversed sothat the group that had the first com-puter lesson now gets the lecture by tra-ditional means and the other group getsthe computerized material. Each half ofthe class will be tested by the same ex-amination. This type of design will ob-viously allow the students to be theirown controls and a comparison of thelecture versus the computerized materialcan be made for the entire class.

One of the most attractive thingsabout CAI to the Physiology Depart-ment concerns the available literaturetoday which seems to indicate that theless endowed students receive the mostbenefit from the technique. The stu-dents at the top of the class do about aswell regardless of the method of mate-rial presentation. It is the feeling of thePhysiology Department that the com-puter could, at the very least, provide asupplement to the traditional education-al system which would be helpful to thestudents having difficulty with the tradi-tional system.

In the Department of Anesthesiology

103

there is a lesson developed from a paperauthored by Dr. Joseph M. White andDr. James A. Cutter concerning resusci-tation and anesthesia in crushed chestpatients. This is an approach to the useof computer assisted instruction in thecontinuing education of physicians. Thematcrial is in a tutorial mode and ex-plores the latest techniques in emergen-cy cases concerning resuscitation andanesthesia. The method of implementingthis lesson involved the direct program-ming of a paper which appeared in theJOURNAL OF AMERICAN SURGERYentitled "Resuscitation and Anesthesiain Crushed Chest Patients". The lessonhas been presented to a group of 16 gen-eral practitioners who attended a two-week course at the University of Okla-homa Medical Center in postgraduateclinical anesthesia. Incorporated in thecourse is the possibility of selecting anyarea of the subject the student wishes toreview by simply typing "outline". Thiscauses the outline of the material cov-ered in the lesson to appear on thescreen. This capability is available to thestudent at any point during the lesson.The lesson has also been programmedfor the word "Comment" to be recog-nized by the computer anytime it is en-tered by the student. The screen isimmediately cleared and the student isallowed to type whatever he likes. Hismessage is recorded on magnetic tapeand the program subsequently branchesback to the point reached in the lessonwhen "Comment" was requested. Thisgives the student the ability to disagreewith any portion of the lesson or tomake comments to the author at anytime. There are strong indicators at thispoint that computer assisted instructionwill be a significant factor in the futureof continuing education of physicians.Many of the traditional problems, suchas distance of the physician from alearning center and lack of time for

extensive blocks of continuing educationeffort, are transcended by the very na-ture of computer assisted instruction. Itis entirely conceivable and practical atthis point for a terminal to be placed ina remote station and tied by voice gradetelephone lines to the computing facilityin a learning center such as the Univer-sity of Oklahoma Medical Center.

Certainly one of the requirements ofthe successful development of a comput-er assisted instruction effort is that thefaculty understand and want the facili-ty. These requirements are met at theUniversity of Oklahoma Medical Center.The faculty is not only interested butthey are also permissive in allowing theother disciplines (computer theorists andlearning theorists) to develop their tech-niques which are fundamental to an ef-fective system. An effective marriagemust exist between the computer scien-tists, the learning theorists and the fac-ulty. Real success can ultimately be real-ized only when these individual effortsare integrated in an environment of effi-cient and effective learning for the stu-dent.

SummaryThe following is a summary of the

CAI material made available to the con-ferees during their visit to the Universi-ty of Oklahoma Medical Center:1. THE CARDIOVASCULAR SYS-

TEM (School of Health ) A drillsession concerning basic cardiology.This is an excerpt from the lesson onthe cardiovascular system presentedto the graduate class in MedicalBackgrounds taught by Dr. ThomasN. Lynn.

2. THE HUMAN SPINAL CORD(School of Health) A tutorial ap-proach to the teaching of the basiccomponents and functions of the hu-,man spinal cord. This is an excerp:

3.

4.

5.

6.

7.

from a lesson by Dr. Thomas N. Lynnfor the Medical Backgrounds class.INFANT NUTRITION (Pediatrics)

Two lessons concerning feedingprinciples and infant nutrition in atutorial mode were prepared by Dr.Arthur W. Nunnery of the Depart-ment of Pediatrics. One lesson dis-cusses breast feeding and the other,formula preparatidn. The material isused by junior and senior medicalstudents.ACTION POTENTIAL (Physiolo-gy ) A tutorial exposition concern-ing the action potential in Physiology.This lesson is being used by freshmanmedical students. The professor is Dr.Roger Thies of the Department ofPhysiology.PANCREATIC SECRETION (Phy-siology) A tutorial approach to theteaching of pancreatic secretion byDr. Eugene D. Jacobson of the De-partment of Physiology. This lesson isis being usied by freshman medicalstudents.SAMPLING THEORY ( Biostatis-tics ) This lesson is an integratedapproach utilizing tutorial and prob-lem solving techniques in computeraided instruction. The professor isDr. Robert C. Duncan of the Biosta-tistics and Epidemiology Depart-ment. The lesson is to be used bygraduate students in the Departmentof Biostatistics and Epidemiology.RESUSCITATION IN ANESTHE-SIA AND CRUSHED CHEST PA-TIENTS (Anesthesiology) Thislesson is designed to explore the pos-sibilities of the use of CAI techniquesin research reporting. The materialwas authored by Dr. Joseph M.White and Dr. James A. Cutter. It isenvisioned as an approach to the useof computer aided instruction in thecontinuing education of physicians.

CLOSING COMMENTS

This conference has been concerned with the role of computers in medical

education. Past and present experiences of knowledgeable leaders have been

presented and discussed as well as new theories and ideas. Plans for the future

solution of problems and expansion into operational services to education have

been expressed, and it appears that the tools are available to meet the challengespresented here. Our ability and desire to do the job remain to be garnered.

Two resolutions were presented and endorsed by those in attendance, and

they are reproduced here.

1. "The heart of this conference has been interaction among knowledgeablepeople. It was the first of its type in this respect and dealt with an im-portant topic, computer aided instruction in medical education. Be itresolved that a conference utilizing the interactive format and discussingcomputers in medical education be held on an annual basis."

2. "Be it resolved that we commend the University of Oklahoma MedicalCenter and the United States Public Health Service for their conduct andsupport of this conference and in particular that we commend Dr. JamesL. Dennis, Dr. Joseph M. White, Dr. Edward N. Brandt, Jr. Mr. WilliamHarless and Miss Elizabeth Ray of the University of Oklal;oma MedicalCenter, and Dr. Frank McKee, Mr. Norman Tucker and Mr. NormanRydland of the United States Public Health Service and the most ef-ficient supporting staff."

The conference was adjourned at noon on Friday, April 5, 1968.

105

104

document resume - eric · education. five speeches and four panel discussions dealt with computers....

Documents