LIBRARY
OF THE
MASSACHUSETTS INSTITUTE
OF TECHNOLOGY
ALFRED P. SLOAN SCHOOL OF
THE IMPACT OF COMPUTERS ON KNOWLEDGE INDUSTRIES:
PART II*
Charles A. Myers
420-69
October 1969
MASSACHUSETTS
TITUTE OF TECHNOLOGY
50 MEMORIAL DRIVE
IDGE. MASSACHUSETTS
oci 8 lyw
DEWEY LIBt^ARY !
THE IMPACT OF COMPUTERS ON KNOWLEDGE INDUSTRIES:
PART II*
Charles A. Myers
420-69 October 1969"
The research and release time for this study were supported
in part with funds from the Inter-University Study of Labor Problems in
Economic Development and from a research grant on the implications of
technological change on organizations made by the Ford Foundation to the
Alfred P. Sloan School of Management, M.I.T. I am indebted to Ephraim
R. McLean 3rd for his conscientious assistance in the research for this
study during 1967-68 and subsequently as he reviewed drafts.
Part I, Sections 1, 2 and 3, Working Paper 392-69, May 1969.
Comments are welcome on this first draft of Part II.
. A/ Q j/i
/Uj,
I N0\/ rj 1969
M. I. I. LIBkAKItS
II-l
Medical and Hospital Services
As in the other areas examined in the preceding sections, the
application of computers to medical and hospital services is moving
rapidly. Keeping up to date on developments in this field is often as
difficult as it is in the other "knowledge industries." An editorial
in the special 1966 issue of the Journal of the American Medical Asso-
ciation on computer applications posed the question:
"What is the present status of the computer in medicine?
The honest answer to that question is, of course, that we
do not know. In common with publications in all fields
of human knowledge, medical literature is old the day it
appears in print. "^
In attempting to assess what has happened, it is important to
remember the caveat suggested by Dr. G. Octo Barnett, Director of the
Laboratory of Computer Science at the Massachusetts General Hospital in
Boston:
"I am concerned with the scientific integrity of the approach
that allows a broad statement that the computer can do tasks
X, Y^ and Z when even a superficial investigation would reveal
that X can be done only on a limited demonstration basis in an
artificial environment, Y is being seriously considered as an
area to be programmed three years hence when system ' . . . ' is
fully functioning, and task Z looks like a challenging problem
in information processing and wouldn't it be nice if we could
get a National Institute of Health grant to carry out a research
project. The failure to discriminate between present reality and
JAMA, vol. 196, no. 11, June 13, 1966, p. 1014.
534428
II-2
future speculations has been one of the major causes of frus-
tration and misunderstanding in the medical applications of
computer science .
The Needs
A review of the literature on computer applications in medical
and hospital services indicates the following needs which have called
forth these efforts:
1. The information explosion represented by increased records for
present and past patients means that professional hospital personnel may
spend a third (or more) of their time on information processing. Much of
this is routine clerical work with manual methods which have been used
2
for many years. The consequences of the record-keeping explosion are
dramatically shown in the experience of the Massachusetts General Hospital:
"The Medical Records Department now stores 1,350,000 records
dating from 1937. During an average day, 150 new patient
records are issued and over 4,000 demands for existing records
are made. ... The various laboratories perform 1,000,000 pro-
cedures per year... Each time a patient is admitted, pertinent
information is sent to some 30 different areas. It is estimated
that on an average day, some 5,000 doctor's orders are written
and the nursing service administers over 30,000 drugs and treat-
ments. It is probable that at least 50,000 separate items of
information are entered into patient records each day, or almost
20,000,000 separate items each year. The problems created by the
large bulk of information processing are greatly magnified by the
In Report to the Computer Research Study Section , Research Grants Review
Branch, Division of Research Grants, National Institutes of Health, on
Computer Applications in Medical Communication and Information Retrieval
Systems as Related to the Improvement of Patient Care and the Medical Record,
September 1966. (Also contains an annotated bibliography.) (The under-
lining added above.)
2
"It is probably true that many of the record keeping methods we use (in
hospitals) are somewhat archaic and represent the carry-over of practices
developed 30 to 50 years ago." Jordan Baruch and G. Octo Barnett, quoted
in Datamation , December 1965, p. 30.
II-3
complexity of the data flow. For example, there are 24 differ-
ent laboratories in the MGH which perform over 300 different
laboratory tests involved in routine patient care. ... The
manual transformation from the patient domain (the doctor's
order book) to the time domain requires many hours of clerical
work on the part of the nursing staff. For example, a nurse
in each care unit may spend 4 hours each day on bookkeeping
activities concerned with the ordering, administering and
recording of medications."^
2. There is a chronic shortage of all kinds of medical and hospital
personnel, and this has recently been accentuated with the advent of
Medicare. One estimate is that "with our continuing population growth,
2
333,000 physicians will be needed by 1975." Perhaps even more serious
is the shortage of nurses, medical technologists and other types of para-
3
medical personnel. To some extent the lower salaries, long hours, night
shifts and other conditions of employment which are less favorable than
those available in other skilled occupations may account for the "shortages.'
Whatever the reasons, the existence of unfilled vacancies in hospitals
and medical centers continues. Research studies of the health manpower
field are proceeding in a number of places, and "action programs" are
proposed at federal and state levels. New medical schools are being
established. But present shortages are not likely to be alleviated within
the foreseeable future.
G. Octo Barnett, The MGH Report, Hospital Computer Project, February 1966,
pp. 14-17.
2
Evon C. Greanias, "The Computer in Medicare," Datamation , vol. 11, no. 12,
December 1965, p. 25. However, another projection concludes that there will
be 362,000 physicians by 1975 - hence no shortage, cf. Rashie Fein, The
Doctor Shortage , Brookings Institution, Washington, D. C, 1968.
Technology and Manpower in the Health Service Industry, 1965-75 , Manpower
Research Bulletin No. 14, U. S. Department of Labor, May 1967, especially
pp. 18-24.
II-4
3. While the practice of medicine has required more sophisticated
information in greater quantities, professional medical personnel have
less time to process it because of the shortages mentioned above. Thus,
the possibility of human error is probably increased. One hospital
reported that there were 5 per cent errors in all of the manual process-
ing of the administration of medication through the hospital pharmacy.
It has been claimed that even the doctor's memory needs an "electronic
crutch" through computerized medical information retrieval or through
computer-aided diagnosis, discussed below. The information explosion
has swamped the medical profession, sometimes with serious consequences.
The workload of lab technicians in hospitals is increasing each year, but
they must spend as much as 2 hours daily on clerical work, with possible
errors in transcription. Thus, the opportunity for computer-based informa-
tion processing is present, although the task of developing a computerized
2
system is never easy. Also, with the present more-powertal drugs, mistakes
can be more costly. Using the wrong drug or the wrong amount is morp likely
to be fatal than formerly.
According to a New York Times editorial, "Many people die needlessly of
Hodgkins Disease, a cancer of the lymphatic system, because their doctors
are ignorant of the major advances in treatment made during recent years.
. . . Presumably there are many other diseases in which the gap between the
most effective treatment known and that actually received by many patients
is very wide indeed. One reason is that even physicians who seek to keep
up with the advances in medicine are swamped by conflicting demands of an
overly busy practice and the information explosion in their field." New
York Times . February 19, 1969.
2
Two doctors reported in 1965 that one such computerized system was discon-
tinued because of errors in input, excessive time requirements on the medical
residency staff, unacceptable time lags, and, above all, a cost much greater
than originally estimated. H. W. Baird and J. M. Garfinkel, "Electronic Data
Processing of Medical Records," New England J o urnal of Medicine , vcL 272, 1965,
pp. 1211-1215.
II-5
4. Finally, the process of diagnosis itself is time-consuming.
Patients may not immediately be referred to the right specialist, with
consequent loss of time to doctor and patient. Then the actual diagnosis
may take much longer, if tests are spaced out and if a number of specialists
have to be consulted. The administrator of the Bellevue Hospital in New
York City is quoted as saying: "If you're acutely ill, we can handle you.
But if you've got diagnostic problems, it can take six weeks to get a
• ..1
diagnosis."
First Steps
The first computer-applications reported in the literature cover
medical records, hospital administration, electrocardiagram analysis, and
medical research. All of these are continuing, so that distinctions be-
tween first steps and recent applications are somewhat hazy.
1. Medical records . By 1965 about 200 large hospitals had either
installed computers systems or had computers on order for storing some medical
2
records and for other administrative purposes. Some of these were
probably "experimental," and research-oriented, as was one of the most
publicized: the Massachusetts General Hospital project under the direction
Dr. R. A. Wyman, quoted by Dr. John F. Davis, "Computers and Medicine,"
International Science and Technology , December 1966, p. 46.
2
Technology and Manpower in the Health Service Industry . 1965-75, p. 46.
For an account of conversational computer interviewing of medical patients
to obtain medical histories, see Werner V. Slack, G. P. Hicks, C. E. Reed
and L. J. Van Cura, "A Computer-Based Medical History System," New England
Journal of Medicine , January 27, 1966. Dr. Slack is Professor of Computer
Science and Medicine at the University of Wisconsin. The system is now
reported to be in operation, using a cathode-ray tube for showing questions
to the patient, who responds by pressing certain buttons.
II-6
of Dr. G. Octo Barnett. It began in 1962 with the assistance of Bolt,
Beranak and Newman, a Cambridge consulting firm. The purposes of this
project initially were to (1) "increase the rapidity and accuracy of
collecting, recording, transmitting, retrieving and summarizing patient
care information," (2) "decrease the amount of routine paper work required
of the nursing staff," (3) "arrange and consolidate information for more
effective and efficient utilization by the medical staff, and (4) "store
large amounts of complex medical information and contribute to clinical
research by facilitating rapid and easy retrieval and analysis of stored
information."
The first MGH "developmental-experimental project was the use
of a small time-shared computer to process drug orders in the hospital's
pharmacy. The computer program includled correct formulae, dosages and
other directions, so that orders transmitted by doctors or nurses were
checked for accuracy before being filled. Doctors or nurses could over-
ride the dosage limits, but only on notification by their names. However,
the former non-computerized system continued in parallel with the
new system, with the result that the staff withheld "full support and
utilization of the experimental system in favor of the old-fashioned,
2
time-tested one."
MGH Project Status Report, February 1, 1966, p. 1.
2
MGH Project Status Report, February 1, 1966, pp. 24-25. The total staff
of the original project numbered 71, 34 of whom were from the hospital
and 37 from the consulting firm. It was financed by grants from the
National Institutes of Health, the American Hospital Association, and
the American Heart Association. This pharmacy application is now running
successfully at Monmouth Hospital, Monmouth, New Jersey.
II-7
The computer project at the Tulane University School of Medicine,
directed by Dr. James W. Sweeney, began pilot studies in 1961 looking
toward electronic storage and retrieval of patients' medical records. These
included records for patients in cardiac surgery, cancer detection, obstetrics,
gynecology, orthopedics, pediatric cardiology, and psychiatry. By 1966 there
were more than 40,000 individual case histories in Tulane' s computer data
bank. Information was submitted by physicians on self-encoding worksheets
for each patient. To make subsequent information retrieval for the physi-
cian easier, a natural language (Meditran) was developed to permit the
physician to use words to specify what information he wanted.
2. Hospital administration . Accounting, payrolls, and dietary
administration have been natural computer applications in hospitals and
medical centers. The Minnesota Hospital Service Association provides a
computerized service to 180 hospitals for all of these purposes, and
2
eventually for medical records. Other groups of hospitals in New Jersey,
and in Peoria, Falls Church, and Wilmington, among others, are sharing
computers for accounting purposes. Patient billing has also been speeded
up by computerized systems. The Abbott Hospital in Minneapolis, for example,
draws on the central file mentioned above for patients' records, and gives
the patient his bill "in a matter of seconds, whereas formerly it took
considerable time for the office staff to look up all the individual charges."
"Packaging Patient Information," Hospital Physician , September 1966, pp. 45-46.
"Hospitals Share Computer Through Communications," Systems , February 1967, pp.
12-14. Hospital administration is apparently still the major application of
computers, despite the interest in more recent experimental applications. See
J. Peter Singer, "Computer-Based Hospital Information Systems," Datamation ,
May 1969, pp. 38-45. This is an up-to-date survey of recent developments.
•^ Ibid. . p. 14
II-8
3. Electrocardiagram analysis - George Washington University
Hospital In Washington, D. C, began In 1963 to develop a computer pro-
gram for analyzing abnormalities and deviant patterns In electrocardla-
grams. The Veterans' Administration began a similar effort before 1965.
The Mt. Slnal Hospital in New York programmed a computer to recognize and
Interpret rhythm changes in the heart's beat^ before 1966.
4. Research As In other types of scientific research, the computer's
facility for rapid data processing has relieved medical research workers
of much of the drudgery in their work and has speeded "critically important
2
experiments." As noted earlier, a number of the early computer applica-
tions in medical and hospital services could also be termed "research."
Recent Applications
Some of the initial projects have continued; these and additional
computer applications will be summarized under the following headings:
diagnosis and treatment, information retrieval systems, planning and
managing complex medical and surgical procedures, hospital administration,
and medical education.
1. Diagnosis and treatment . In an early summary of computerized
medical diagnosis, four general areas were listed as relevant: (a) "the
communication of Information about the patient to the physician, (b) com-
parison of patient information with available medical information,
As reported variously in Datamation , December 1965; Technology and Manpower
in the Health Services Industry , 1965-75, p. 40; and New York Times , March 3,
1966.
2
Business Week , July 9, 1966, p. 142. This article has a summary of early
use of computers in medical research laboratories.
II-9
(c) diagnostic decision making, and (4) treatment of the patient." A
more visionary and optimistic view of the future was outlined by Evon C.
Greanias of IBM:
"Diagnostic assistance by computers is an attractive
potential application. ...In a typical exchange with the
computer, a physician might type in a few symptoms, and the
computer would respond by printing out a list of possible
diseases. The physician might then ask why a particular
disease appears on the list, and the computer would supply
a number of possible cause-and-ef feet relationships. The
physician could then ask for more information about a
particular causal path, examine the bgic, and agree or
disagree with the computer's response. Searching in this
way, the doctor could make a logical examination of all the
reference possibilities stored in the machine. "^
Despite this optimistic view, a 1966 report concluded that
"there has been little operational success in the area and most of the
3
work has been either theoretical, developmental or sharply limited."
Among the more recent attempts to develop computerized systems
for medical diagnosis are the following:
« Preliminary diagnosis has been used in periodic medical exams
for preventative medicine. The Kaiser Foundation Health Plan
in the San Francisco area gives the patient a self-administered
questionnaire of 266 questions; answers and routine test results
are then fed into the computer. This information can then be
Lee B. Lusted, "Computer Techniques in Medical Diagnosis," Ch. 12 in
Ralph W. Stacy and Bruce D. Waxman, Computers in Biomedical Research ,
vol. I, Academic Press, New York, 1965, p. 321.
2
Datamation . December 1965, p. 27.
3
Dr. G. Octo Barnett in his MGH Project Report, September 1966, Appendix,
p. 25. The one exception to this generalization was, he noted, "the
automatic classification of electrocariographic abnormalities" by computers.
11-10
used by the physician in his subsequent examination of the
patient. This experience seems to be the prototype of the
concept of "automated multi-phasic health screening (AHMS)"
which is claimed to streamline the "time-consuming, office
bound physical examination" by reading, processing and
interpreting past medical records, analyzing other tests,
and producing a printed report for the company or family
2
doctor.
The Mayo Clinic has used a computer analysis of the results
of the Minnesota Multiphasic Personality Inventory, to provide
the doctor who is not a trained psychiatrist or clinical psy-
chologist with a clinical evaluation of the test results on a
particular patient. The computer center at the Rockland State
Mental Hospital in New York "enables psychiatrists to maintain
accurate, standardized medical histories; someday they may also
help the diagnosis." The system includes pencil notations on
forms that describe a patient's behavior and the computer then
stores this information to provide the psychiatrist later on
on call with a full patient history in straight sentences.
The Lahey Clinic in Boston is also using this system.
As reported by Paul Armer in Appendix Vol. 1, p. 225, Report of the
National Commission on Technology, Automation, and Economic Growth,
Washington, D. C, 1966.
2
"Electronic First-Aid for the Busy Doctor," Business Week . October 5,
1968, pp. 156-158.
3
Lusted, op-cit., p. 322. Also see, Howard P. Rome, M.D. , William C.
Menninger Memorial Lecture, "Psychiatry: Circa 1919-1969-2019,"
American College of Physicians, Spring Meeting, Chicago, April 25, 1969.
4
Thomas Fleming, "The Computer and the Psychiatrist," New York Times
Magazine . Sunday, April 6, 1969, p. 47.
II-ll
Computerized collections of medical histories are being developed
at the Lahey Clinic and the Massachusetts General Hospital in Boston for
preliminary diagnosis and patient assignment to physicians. Professor
John F, Rockart of the M.I.T. Sloan School of Management has been working
with the Lahey Clinic on the development of a "symptom questionnaire" sent
to patients prior to their visit, to determine which specialists they should
see, thus saving valuable time during consultations. A computerized sys-
tem is in the "pilot plant" stage of being used to schedule patients more
efficiently through the multi-specialist clinic. Cost effectiveness and
2
attitudinal studies are being made of these experimental programs. The
MGH is also working on an automated collection of medical histories of
patients in the ambulatory clinic, including automatic ordering of partic-
ular laboratory gests, and minimal automatic differential diagnosis of
those histories specifically oriented toward a complaint, such as a chest
pain, a headache, etc.
John F. Rockart, Philip I. Hershberg, Jerome Grossman and Richard Harrison,
"A Symptom- Scoring Technique for Scheduling Patients In a Group Practice,"
M.I.T Sloan School of Management Working Paper 385-69, to be published in
the Proceedings of the Institute of Electrical and Electronic Engineers .
2
Report of the Activities of the Laboratory of Computer Science, Department
of Medicine, Massachusetts General Hospital, 1968. Bolt, Beranak and Newman
is no longer involved with this project, which has developed into a series
of several smaller, better-focused research studies. Dr. Barnett observes:
"We have abandoned the 'total systems' approach to the development of a
hospital information system in favor of a modular evolutionary activity,
wherein we proceed in a type of hill-climbing fashion, developing and
implementing relatively separate sub-systems. We feel that this is a much
more productive approach and that integration of the sub-systems will be
relatively simple compared to the task of the development of each sub-system,
p.l
11-12
A different approach to computer-aided diagnosis is being
developed at the Massachusetts General Hospital by Professor
G. Anthony Gorry of the Sloan School of Management in collabo-
ration with Dr. G. Octo Barnett, Director of the Laboratory of
Computer Science at the hospital. Their approach is known as
"sequential diagnosis" to correspond to the sequence of steps
a doctor uses in diagnosis, as opposed to present computer
diagnostic systems which start out with all the relevant data,
including tests which may be unnecessary and costly. The
sequential system involves three parts: (1) the information
structure constituting the medical experience available to the
program, including probabilities of certain diseases occurring
with certain symptoms, and costs of various tests and mis-
diagnoses; (2) the inference function which uses Bayes rules
to update a probability distribution for the disease in ques-
tion, as the sequential interaction of the doctor with the
computer in English conversation proceeds in real time; and
(3) the test selection function, which selects one of several
decision alternatives corresponding to each potentially useful
test and one corresponding to the cessation of testing. This
step takes into account the current view of the problem (as
seen from the inference function), the cost of the test, and
G. Anthony Gorry and G. Octo Barnett, "Sequential Diagnosis by Computer,"
Journal of the American Medical Association , vol. 205, September 16, 1968,
pp. 849-854; and G. Anthony Gorry, "Modelling the Diagnostic Process,"
M.I.T. Sloan School of Management, Working Paper 370-69, February 1969.
A brief account of this approach was also reported in The Technology Review ,
April 1968, p. 51.
11-13
possible test results. The test selection function determines
the best test to perform at a particular stage in the diagnosis,
as contrasted with other computer-aided diagnostic programs
which begin with a number of possibly-helpful and often costly
tests.
This sequential diagnostic program has been employed in
the diagnosis of bone tumors, congenital heart disease (35
different types), and recently, acute renal failure. These
programs have been developed on M.I.T.'s pioneer time-sharing
computer, Project MAC, preliminary experience supports
the superior value of sequential-decision making in computer-
aided diagnosis, although the researchers state that more
extensive evaluation will be necessary. One additional
value of the system may be that para-medical personnel such
as nurses could query the program and know when a patient
should be referred to a specialist on the disease.
Professor Gorry is also working with the New England
Medical Center of Tufts University in developing similar
computer-aided sequential diagnostic programs in collaboration
with doctors who help revise and test the programs. These may
later be available to doctors throughout the region to assist
diagnosis of certain diseases.
t Computer-aided electrocardiagraphic analysis of heart
patients is being extended over wider geographic areas. Early
in 1969, a computer at Mt . Sinai Medical Center in New York City
^Gorry and Barnett, op. cit. , p. 054. Professor Gorry is extending the renal
failure study to deal with the assessment of therapeutic strategies when
significant risk is involved.
11-14
was available to doctors in West Virginia through a project
carried out jointly by Mt. Sinai, the West Virginia State
Health Department and the Cro-Med Bionics Corporation. The
program is still experimental, since the EKG readings trans-
mitted to the New York computer by special modulated signals
over the telephone are being checked by mailed magnetic tapes
and by independent EKG analysis. In the past. West Virginia
hospitals have waited two weeks for an analysis by distant
cardiologists, and the computer-aided system promises two-
minute results by telephone to the waiting doctors.
* A computer has been teamed with a device called a scintiscanner
by doctors at Long Island Jewish Hospital to find thyroid and
brain tumors directly in patients, and to pinpoint the location
and size of the growth. The computer analyzes the x-rays, by
a technique derived from military uses to examine aerial photo-
graphs for enemy troop concentrations. The computer can detect
shades of gray in the x-ray, indicating tumors which visual
2
examinations may miss.
1
"Computer to Aid Heart Diagnoses," New York Times , December 8, 1968,
p. 59.
2
"Computer Helps to Detect Tumors," Boston Sunday Herald , March 2, 1969,
Section A, p. 3.
II-14A
On-line computer monitoring of heart and lung conditions of
patients in intensive care units has been developed at the Pacific
Medical Center in San Francisco, as a joint project with IBM. It is
claimed to be more than automated record-keeping, because it "is aimed
at spotting potentially dangerous conditions early enough in their
development cycle to correct hem easily and minimize their medical
effect on the patient." The system, which first went into operation
several years ago, also provides for the rapid access to data needed
by medical personnel to take the corrective action. Skin temperatures
and blood pressure are also monitored by the system.
James 0. Beaumont, "On-Line Patient Monitoring System," Datamation ,
May 1969, pp. 50-55.
« A matrix-type diagnostic computer has recently been patented
in the United States by the Nippon Electric Company Ltd. of
Tokyo. According to a published account, a similar computer
has been used for some time in the Tokyo University Hospital
and 50 other Japanese hospitals. Symptoms are linked with
diseases in the computer's memory, and weight is given to the
degree of certainty associated with symptoms for a particular
disease. When buttons are pressed on the matrix, "an ammeter
indicates the most likely disease."
• Medical diagnosis is also claimed to be aided by a Clinical
Decision Support System developed before 1967 by Evon C. Greanias
2
(an engineer) and Dr. Frederick J. Moore of IBM. A more recent
advertisement by IBM stated that Dr. Moore was experimenting
with the system which would make available to doctors' offices and
even hospital rooms "the latest information about ailments and
their remedies... Using a portable terminal no bigger than a
briefcase with a keyboard and display screen, the doctor could
get the information he wanted by simply plugging into the circuit.
"Japanese Company Promotes Diagnosis by Computer," New York Times ,
August 31, 1968, p. 42. The computer has circuits for AO diseases and
40 symptoms. It is obviously quite different from the "sequential
diagnosis" developed by Gorry and Barnett, especially since the Times
report further states: "The company says its operation does not re-
quire professional knowledge of medicine or electricity; the patient
himself could push the buttons."
2
New York Times . September 25, 1967; see also Datamation , December 1965,
p. 28.
11-16
With the latest in medicine at the doctor's fingertips, combined
with his judgment and experience, he could make his diagnosis
in the shortest possible time."
This approach, which involves "InfotTnation retrieval" as
an aid to diagnosis, would apparently not be interactive, and is
simply designed to help the "already overworked M.D." overcome
the medical information explosion.
2. Information gathering and retrieval systems . Early efforts
to gather medical records and other medical information for com-
puter storage and retrieval have been mentioned earlier, and some
have been reviewed in the preceding section. Among the more recent
developments are the following:
• The laboratory test reporting system, which is the largest
on-going research project of the Laboratory of Computer Science
is claimed to be
at Massachusetts General Hospital, and/the largest of its kind in
any hospital in this country. The project began with the Chemistry
Laboratory, with direct reporting of test results to the emergency
ward and intensive care units. A later stage of the project will
provide for direct connection to the automated instrumentation in
the Chemistry Laboratory, and eventually extension of the computer-
aided reporting system to the other 32 laboratories in the hospital.
Later on, the system can be transferred to other medical centers
for use in their laboratories.
Report of Activities of Laboratory of Computer Science, 1968, pp. 1-2.
# A medical information network, for hospitals is the objective
of Medinet, a General Electric subsidiary centered in the Boston
area. By early 1969, this system was being tested in a number of
hospitals in New England, and was actually in use in one hospital.
• The National Library of Medicine in Bethesda, Maryland, has
been perfecting its MEDLARS (Medical Literature Analysis and
Retrieval System), developed initially in 1963. Its intention
was to produce a monthly computerized list of some 14,000 articles
in the bio-medical field, but delays were reported by physicians
in utilizing it effectively.
• A computer-based Medical Audit Program (MAP) and Professional
Activity Study (PAS) are both available to hospitals through the
Commission on Professional and Hospital Activities, Ann Arbor,
Michigan. By the end of 1967, approximately 1,140 hospitals were
involved in the PAS system, which covered primarily hospital admin-
istration and research. Promotional literature described the
systems as "computer-based medical records information systems,"
and the non-profit sponsoring organization was described as a
2
computer utility" service for hospitals.
• Some writers have envisioned a medical information system
similar to the "total management information system" in industry
(which does not yet exist). One notes that this is further in
Hospital Physician . September 1966, p. 48. See also New York Times ,
February 6, 1966, and Guide to Melars Services, National Library of
Medicine, Section F, p. 1, U. S. Department of Health, Education and
Welfare, November 1966.
2
Correspondence and enclosures from Berhnard J. Henehan, December 13, 1967,
11-18
the future, and another and another has abandoned the effort in
2
his own hospital in favor of a subsystem by subsystem approach.
A jjroposed national computer-based medical record system has been
discussed in Great Britain, which has a national health service ;
and the Food and Drug Administration in the United States claims
to have put a major computer system in operation to give all its
experts in field offices better access to the agency's technical
data. It would provide instant data on new drugs under investi-
gation, and would monitor abuses of narcotics and psychedelic
drug compounds.
3. Planning and managing complex medical and surgical procedures .
Using Critical Path and PERT methods taken from computer applications
in industry, a team of doctors have planned and carried
out kidney transplant operations through a number of computer runs.
Simulation of procedures in advance of the actual operation would
seem to be a logical application of computer-based systems, but
published reports on further applications are lacking.
Evon C. Greanias in Datamation , September 1965, pp. 27-28.
2
G. Octo Barnett, 1968 Report as cited, p. 1.
^"U.K. Council Proposes Medical Data Bank," Datamation , September
1967, p. 105.
"Drug Computer to Expand Data," New York Times . August 12, 1967,
p. 11; also Business Week , August 19, 1967, p. 60.
"Planning and Managing Complex Medical and Surgical Procedures," in
special issue on computers. Journal of the American Medical Association ,
vol. 196, no. 11, June 13, 1966.
11-19
4. Hospital administration . In addition to accounting, payrolls,
patient billing, dietary administration and menu planning (all noted
earlier), some recent applications have involved various forms of
"inventory control" such as:
• Assuring full bed occupancy, by careful patient scheduling as
has been done at the Massachusetts Eye & Ear Infirmary.
• Control of blood bank supplies, to assure use in right order
to prevent spoilage and losses.
5. Medical education . As computer experiments and applications
in medical and hospital services are spreading, it is natural that
they should be introduced in medical eduation. Some examples are the
following:
9 Fourth-year medical students at the University of California at
Los Angeles are given computer presentations of hypothetical patient
medical histories and texts which they then diagnose and recommend
treatment. The computer program checks the validity of their work.
This is claimed to be superior to clinical experience, in which the
student finds diagnosis and treatment developed by the staff, with
2
little opportunity for his own separate diagnosis.
• A computer-controlled patient simulator has been developed at
the Medical School of the University of Southern California to train
3
resident physicians in anesthesiology.
This was Lockheed Aircraft's first venture into computerized hospital
systems. Fortune. January 1967, p. 186. Lockheed was also developing
a computer center or utility for 58 San Francisco Bay hospitals.
2
New York Times , February 2, 1968.
Datamation , May 1967, p. 77.
11-20
• The Laboratory of Computer Science of the Department of Medicine
at the Massachusetts General Hospital has engaged in a number of teach-
ing efforts, including a "medical student medical record," which uses
a medical summary completed by each student on his cases in the
clinics, to report both to the students and the faculty a summary
"so that they can have a better understanding of the type and charac-
ter of the treatment they give." A lecture- laboratory course in
computer science is also offered to clinical and research fellows
of the hospital. Some Harvard Medical School students are employed
part-time in the Laboratory of Computer Science.
Some Long-Run Implications of Computer Applications in Medical and Hospital Servic
1. There will be further reduction of routine, time-consuming work
by professionals and sub-professionals. If it is remotely true, as one
computer specialist has claimed, that 90 per cent of the physician's time
is spent in information gathering for patient diagnosis and treatment, then
2
this "is the kind of task that can be programmed into a computer." As the
Surgeon General of the U. S. Public Health Service has put it: "The challenge
to the physician is to use computers to help do the burdensome part of the
medical workup, so that he may give his full attention to the creative,
3
human part."
G. Octo Barnett, 1968 Report, as cited, pp. 7-8.
2
Evon C. Greanias (IBM), "The Computer and Medicine," Datamation , vol. 11,
no. 1, January 1968, p. 57.
3
Dr. William H. Stewart, in Datamation, vol. lA, no. 1, January 1968, p. 57.
11-21
2. The shortage of professional and sub-professionals in this
field should be somewhat relieved by computer programs which handle the
routine tasks reviewed in the preceding sections. Furthermore, with the
expanding demand for medical and hospital care, it seems unlikely that
there will be any actual displacement of personnel. Their professional
talents will certainly be better utilized, and some expansion of services
should be possible with a higher ratio of patients to physicians. Computer-
ization of patients' medical histories, preliminary assignments to special-
ists, preliminary diagnosis, and easy access to medical records from the
data bank should be a boon, rather than a threat, to busy physicians.
However, this means that some present work of physicians will be
taken over by computers. Those who now fail to keep up to date on recent
developments may be able to do so with information-retrieval systems simi-
lar to Medlars. Computers may, therefore, even help some to become better
physicians.
3. The view that physicians will not be displaced, even though the
nature of their work may be changed, is challenged by a prediction that
the computer and other highly specialized equipment will be operated by
an army of technicians, who will routinely hook up the patients and maintain
2
the equipment. The large number of physicians, as we know them, will
dwfndle into a small cadre of medical scholars. The authors also point
out that no matter how many physicians and related personnel are trained,
"we cannot keep up with the demand for health services. Technologists,
M. S. Blumberg, "Computers Will Augment Physician's Role, Modern Hospital ,
vol. 106, no. 48, 1966.
2
Edmund J. McTernan (Dean of Health Sciences at Northeastern University) and
Dr. Dean Crocker (of Children's Hospital), writing in the Hospital Physician ,
and quoted in "The Computerized Hospital of the Future Is Visualized," The
Boston Herald , March 23, 1969.
11-22
technicians and aides in the patient-care spectrum are means to extend
the productivity of the physician--the human computer." They add that
the human computer has certain weaknesses - failing hearing and vision,
as well as personal worries which distort perception - while the computer
itself is "always available."
The official American Medical Association view is clearly at variance
with the above prediction:
"If one general conclusion is possible, it is the statement
that no computer will attain an MD degree, nor replace a
single physician. The computer can be an extraordinarily
useful aid, and we should become better acquainted with the
potentialities and limitations of these electronic devices.
But despite the theories of certain enthusiasts, the computer
cannot offset any present or future shortage of physicians.
A current relevant quip states: 'Any doctor who can be replaced
by a machine, deserves to be replaced by a machine.'"^
4. The future possibilities of man-machine interaction in medical
and hospital services are impressive. The sequential diagnosis experiment
being developed by Professor Gorry and Dr. Barnett at the Massachusetts
General Hospital, as reviewed earlier, is an illustration of an inter-
active system between the doctor and the computer, designed eventually
to assist the doctor in his diagnosis in real time. So is the electro-
cardiagraphic analysis provided by the Mt. Sinai Medical Center in New York
City for physicians in West Virginia hospitals. The availability of
patients' medical records from a computer data bank is less interactive,
as is any other form of information retrieval from stored sources. But
all of these are computer-aids to decision making, and represent a true
man-machine partnership, as a 1966 article in the Hospital Physician noted:
Editorial in Journal of the American Medical Association (special issue
on computers), June 13, 1966, p. 1015.
11-23
"For most M.D. 's in computer projects the supreme goal is to
help the physician in his management of patients. They don't
want or expect the computer to tell the doctor what to do, but
they believe it can mark out guidelines and perform specific
services that will directly aid his decision making."^
The same observations may be made about interactive computer systems which
can assist nurses, technicians, dieticians, and office staff in hospitals,
as they make decisions and perform their assigned tasks.
5. Some centralization of organization structures in hospitals is
centers in large hospitals or computer
likely to occur as computer /service centers serving groups of hospitals
replace some of the clerical functions which each participating hospital
previously performed. Examples are accounting, billing, payrolls, and
medical records. Possibly the remaining hospital clerical staffs will
perform somewhat different functions: providing computer inputs and
analyzing computer outputs, rather than keeping and retrieving records
directly. Area-wide or regional medical records will require better
inter-hospital cooperation. Centralization of computer services will
make available certain specialized functions more widely (such as
computer analysis of electrocaidiagrams) , and in one sense, may strengthen
decentralized units in the hospital system of the country.
6. All of these developments may herald "The Coming Revolution in
Medicine," to borrow the title of a recent book by Dr. David D. Rutstein,
Head of the Department of Preventitive Medicine at the Harvard Medical
School. Arguing for a complete reorganization of the delivery of medical
care. Dr. Rutstein believes that computers will act as information clearing
Lead paragraph in "The Computer and Its Effect on Hospital Physicians--
Guidiug Diagnosis and Treatment," Hospital Physician , September 1966, p. 47.
11-24
houses in the diagnosis of disease as well as administrators of a vast
and unified medical network — including feedback mechanisms to control
blood pressure and other physiological functions, perform laboratory
tests, and assist in medical research. He also forsees a reallocation
of physician's duties, some delegated to trained technicians, some per-
formed by machines - leaving the physician to spend his time on problems
that demand his special education, training, and talents.
Resistances to be Met and Overcome
The exciting possibilities of computer applications to meet some of
the needs outlined at the beginning of this Section will not be widely
realized until doctors, sub-professionals and others in hospital and
medical care understand how these can help them in their work. The
conservatism of the medical profession is well known, and as one doctor
active in computer programs has noted, this is "another factor which has
limited the initiative of hospitals in the dynamic application of computer
2
techniques to the area of patient care."
If the lessons from both unsuccessful and successful computer appli-
cations can be summarized, they suggest the following steps are necessary
for forward progress:
1. Initial commitment on the part of the hospital administration
that experimentation and eventual implementation are necessary.
2. Direct involvement of the medical and nursing staff, as
well as other specialized groups, in developing, testing, and implementing
David D. Rutstein, The Coming Revolution in Medicine , The M.I.T. Press,
Cambridge, Massachusetts, 1968.
2
Report to the Computer Research Study Section, ...National Institutes
of Health, op. cit. , p. 33.
11-25
subsystems in particular parts or functions of the hospital. Some of
these should be encouraged by hospital management to volunteer for (or
to be assigned to) these projects.
3. Avoidance of what someone has called "a priesthood of
programmers" through which all changes in information handling must go,
is imperative. The same comment could be made about systems designers
who fail to consult practicing physicians. One physician reported
participating in "a simple compilation regarding one disease," for which
over 10,000 questionnaires covering about 20 questions were gathered and
turned over to trained programmers who were not, however, physicians.
The results were useless and the study had to be reprogrammed after
physicians had rephrased the questions and methods of approach. The
full facts in this case are not reported, so it is possible that until the
error occurred, physicians may have abrogated their responsibility to
become involved in the program from its inception.
4. Continuing involvement of hospital or medical center manage-
ment is vital in selecting objectives, determining priorities in areas for
study of computer applications, and subsequent support and help as problems
are encountered.
5. Finally, an evolutionary approach is desirable, because
the profession changes slowly and successful applications depend on full
understanding and appreciation of the ways in which computer-based systems
can aid doctors and hospital staffs, rather than threaten them.
Dr. Irving S. Wright, Professor of Medicine at Cornell Medical College,
in his presidential address before the American College of Physicians,
San Francisco, April 10, 1967, as reported in the New York Times, April 11,
1967.
11-26
Section 5
National and Centralized Data Banks
The uses and potential abuses of national and centralized
data banks for statistical and analytical purposes have a recent but
full history. The dialogue has involved high-level committees, two
Congressional hearings, and numerous supporters and critics. At
this writing, there is no comprehensive National Data Center,
despite the committees' recommendations. But there are several
partial national data banks in a number of Federal agencies in the
Departments of Agriculture, Labor, Interior, Commerce, Treasury,
and Health, Education and Welfare, as well as in the Board of
Governors of the Federal Reserve System. And there are several
centralized data banks in cities and counties, primarily for admin-
istrative purposes. Since many of these contain information about
individuals, as would the proposed National Data Center, the issue
of privacy of information has been injected into discussions with
some heat.
1
As noted in the Report of the Committee on the Preservation and
Use of Economic Data, to the Social Science Research Council,
Washington, D. C. , April 1965. This is known as the Ruggles
Report, after the name of its Chairman, Professor Richard Ruggles
of the Department of Economics at Yale University.
11-27
The Needs as Developed in the Proposal
The need for a National Data Center was developed in two out-
standing committees, supported by a report of the Joint Economic Committee
of the Congress, and subsequently by a committee of the Division of
Behavioral Sciences of the National Research Council. The needs that
such a center would fulfill may be summarized in the following points:
1. The present organization of Federal statistical information,
much of it on machine-readable tapes, does not lend itself to optimal use
of these vast amounts of data for economic research and analysis. (The
Ruggles Report)
2. There is need for a "large scale systematic demographic,
economic and social statistics file" the purpose of which is "the assembly
of statistical frequency distributions of the many characteristics which
groups of individuals (or households, business enterprises or other
reporting units) share." (Testimony of Professor Carl Kaysen at Senate
Subcommittee Hearings; Kaysen was chairman of a Task Force on the Storage
of and Access to Government Statistics, reporting to the Director of the
Bureau of the Budget.)
3. "The data center would supply to all users, inside and outside
the Government, frequency distributions, summaries, analyses, but never
This section draws material from Robert L. Chartrand, "The Federal Data
Center: Proposals and Reactions," Legislative Reference Servi
Library of Congress, Washington, D. C, June 14, 1965; and from testimony
and appendices in the hearings on "Computers and Invasion of Privacy,"
before the Special Subcommittee on Invasion of Privacy, of the House Com-
mittee on Government Operations, July 26, 27, 28, 1966 and "Computer
Privacy," Hearings before the Subcommittee on Administrative Practice and
Procedure of the Senate Committee on the Judiciary, March 14 and 15, 1967,
and February 8, 1968.
11-28
data on individuals or other single reporting units. The technology of
machine storage and processing would make it possible for these outputs
to be tailored closely to the needs of individual users without great
expense and without disclosure of individual data. This is just what
is not possible under our decentralized system." (Kaysen testimony.)
4. The center would include existing bodies of data already
colle^'-ted by such Federal agencies as the Census, the Bureau of Labor
Statistics, the National Center for Health Statistics, the Office of
Education, the Department of Agriculture, and the Department of Commerce.
Additional data generated as a result of the administration of the
federal income tax and the federal social security systems would be
included. But the Kaysen Committee specifically excluded police
dossiers from the FBI, Civil Service personnel records, personnel data
on the armed services, or any other personal information. While no one
could find out an individual's income or tax paid through the Center,
knowledge of aggregates of income, by sources, would be useful to the
Congress and to the Executive Branch in generating tax policy.
5. "The central problem of data use is one of associating numerical
records and the greatest efficiency of the existing Federal statistical
system is its failure to provide access to data in a way that permits the
association of the elements of data sets in order to identify and measure
the interrelationship among interdependent or related observations. This
is true at virtually all levels of use and for all purposes from academic
11-29
model builders to business market researchers." (Edgar S. Dunn, Jr., of
Resources for the Future, Inc., in his evaluation report of the Ruggles
Report, made at the request of the Bureau of the Budget, November 1, 1965.)
6. Better coordination and integration of separate government
statistical programs is essential, and it was the view of the Joint
Economic Committee that the current statistical information is "toally
2
inadequate to meet the changing policy needs of our times."
7. The information explosion requires that behavioral and social
scientists develop computerized information systems before the flow of
research data rises "to blood heights." This was the view expressed in
a committee report of the Division of Behavioral Sciences of the National
Research Council, which is the operating agency of the National Academy
3
of Sciences. The report recommended a decentralized national network
of data banks containing statistical information, and a Federal data
center to coordinate the Government's statistical output.
This report, as well as the full Ruggles Report, is found in the
Appendix to the House Subcommittee Hearings on "The Computer and the
Invasion of Privacy." The Dunn quotation is on p. 255.
2
The Coordination and Integration of Government Statistical Programs ,
Subcommittee on Economic Statistics, Joint Economic Committee, U. S.
Congress, Washington, U. S. Government Printing Office, 1967, p. 4.
3
Communication Systems and Resources in the Behavioral Sciences ,
summarized in New York Times , January 14, 1967.
11-30
The Problem of Privacy
Will a proposed National Data Center expose individuals to loss
of privacy and maybe blackmail by those who gain access to an individual's
records? This concern has monopolized the two Congressional hearings,
and even the experts are divided on the issue. The same problem applies
to the limited national computerized data banks in present federal agencies,
as well as to the county and city data banks, which will be discussed later.
The Critics
An editorial in the New York Times in 1966 put the fears of a
National Data Center in these words:
"Can personal privacy survive the ceaseless advances of the
technological juggernaut? ....The Orwellian nightmare would
be brought very close indeed if Congress permits the proposed
computer National Data Center to come into being. We already
live with the fact that from birth to grave Federal agencies
keep tabs on each of us, recording our individual puny existence,
monitoring our incomes and claimed deductions, noting when we
are employed or jobless, and--through the F.B.I, and similar
agencies--keeping all too close watch on what we say, what we
read and what organizations we belong to.... What is now pro-
posed is the amalgamation of these files, and the creation of
a situation in which the push of a button would promptly dredge
up all that is known about anyone. Understandably, this idea
has brought vigorous protest, in which we join. Aside from the
opportunities for blackmail and from the likelihood that the
record of any single past transgression might damage one for
life, this proposed device would approach the effective end
of privacy. . . . "'■
Vance Packard, the best-seller author who has written on the
erosion of privacy in his book, The Naked Society , has testified before
the Gallagher Subcommittee of the House of Representatives in the same
New York Times , August 9, 1966. It is clear that the editorial writer
was not aware of the more limited proposals summarized in the preceding
section.
11-31
vein, and has criticized the proposed National Data Center for four
reasons: (1) it "threatens to encourage a depersonalization of the
American way of life," (2) it "is likely to increase the distrust of
citizens in their own government and alienate them from it," (3) "a
central file can absorb large batches of data about people but it is
ill-equipped to correct errors, allow for extenuating circumstances, or
bring facts up to date," and (4) it would place "so much power in the
hands of the people in a position of power to push computer buttons."
He adds:
"When the details of our lives are fed into a central
computer or other vast file-keeping system, we all fall
under the control of the machine's managers to some extent
....My own hunch that Big Brother, if he comes to the United
States, will turn out to be not a greedy power-seeker but a
relentless bureaucrat obsessed with efficient (who) could
lead us to that ultimate of horrors, a humanity in chains
of plastic tape." ^
The public statements of Representative Gallagher and Senator
Long, during or following the hearings they conducted, are not dissimilar
from these views. Indeed, Packard quotes extensively from Representative
Gallagher.
The critics from the universities are mostly law professors,
prominent among whom is Alan F. Westin, Professor of Public Law at
Columbia University and author of a book. Privacy and Freedom (1967). In
his testimony before the Long subcommittee in the Senate, Westin expressed
1
Vance Packard, "Don't Tell It to the Computer," New York Times Magazine ,
January 8, 1967, pp. 44-45ff. The last quotation is on pp. 90-92.
11-32
the view that the proposed National Data Center "would clearly lack the
careful development, the system safeguards, the administrative procedures,
and certainly the type of legal framework that would be necessary to pro-
tect individuals and groups if information from many Federal agencies
were to be collected, stored, and used in unspecified ways, thus raising
the very problems of the incorporation of data collections that you
mentioned in your opening statement." Senator Long had expressed fear
that the proposed National Data "Bank" is "designed to store names and
information on citizens so that the simple push of a button will spread
2
a citizen's life history on the computer readout."
In a more recent statement, Professor West in has suggested that
proper safeguards in computerized data banks should include the following:
(1) privacy - whether certain kinds of information should be collected at
all in these systems, the extent to which one piece of information within
a data bank ought to be associated with another, and what information
should be disclosed and circulated and to whom? (2) due process -- giving
an individual the opportunity to know what is in his data file, to challenge
its accuracy, and to contest interpretations based on the facts, and
1
Long subcommittee hearings. Senate, p. 282.
2
"Opening Statement of Senator Long," p. 276. I leave it to the
reader's recollection whether this is what the supporters of the
Center proposed.
11-33
(3) public audit and review, in the form of watchdog or ombudsman- type
agencies, or legislative agencies, or public review boards, etc. Westin
believes that "public concern over the issue is forcing public and private
organizations to become more self-conscious about the role of privacy
and the role of information in our society; and this exercise could— if
properly nurtured and protected — lead us to give more protection to
individual rights than we provided in the pre-computer era."
1
Alan F. Westin, "Computers and the Protection of Privacy," The
Technology Review , vol. 7, no. 6, April 1969, pp. 32-37. Westin's
views are apparently shared by Professor Robert M. Fano, who was
the first director of Project MAC at M.I.T. In a recent unpublished
paper he noted that the Compatible Time Sharing System at M.I.T.
"is far from adequate in its protection of privacy." Although every
user has his own password, he can give this to other users "or to
everybody." Furthermore, "system programmers are fallible, and mis-
takes in the control programs of the system may remain undetected for
a long time. As a matter of fact, they may be discovered by users
inclined to exploit them for their own advantage or for other
malicious purposes. Our experience at M.I.T. indicates that these
dangers are very real and that no community can be assumed to be
immune from them. .. .There is no simple answer to the problem of
protecting a computer system against intruders." ....The right to
privacy must be defined and protected by suitable legislation, and
also appropriate regulations must be enacted to protect the users
of public computer systems." R. M. Fano, "Implications of Computers
for Society," paper presented at the Joint Summer Conference on
"The Computer in the University," Technical University of Berlin
and Massachusetts Institute of Technology, Berlin, Federal Republic
of Germany, July 22-August 2, 1968, p. 8.
11-34
The Defenders and Their Proposals for Protecting Privacy
Most of those who favored the National Data Center for statis-
tical purposes argued that individual economic and social data should not
be revealed, and that the data bank would not contain detailed personal
dossiers about individuals and life histories. Many of these explanations
were offered at the Congressional hearings, frequently in answer to queries
from the subcommittee chairmen or their staffs. However persuasive they
might be to an objective reader of the testimony, they did not convince
the Congressional subcommittees.
After denying that the proposed center would contain any
individual dossier with personal data of any kind, Kaysen subsequently
tried to suggest the following safeguard on revealing other individual
data (such as incomes, job experience, mobility, etc.):
"Everytime anyone calls out a file or a group of files out of
the Data Center, he has to make a record entry that says, 'I am
so and so, I have called out files number so and so, and so and
so, and I have called them out pursuant to such and such a job
order.' Thus, he leaves a trail, and the machine, and the pro-
grams which operate it, can be so organized that nobody can
operate the machine without leaving a trail, unless he tries
to eradicate both the trail and the data in the machine and .
thus shows to his supervisors that something has gone wrong."
Kaysen has also pointed out that the distinction between a personal
dossier and the assembly of statistical frequency distributions based
on individual economic and social data is not self-applying, and
1
Long subcommittee hearings, p. 9. Despite this. Senator Long at a later
point, questioning another witness, said "If we did go on this cradle-to-
the-grave deal, everything in regard to that individual would be collected
and by pushing a button you could get all of the information (on the
individual) if it was put in a machine like that." (p. 116)
11-35
"administrators and bureaucrats, checked and overseen by politicians,
have to apply it. But so it is ever."
The same point was made at the Long subcommittee hearings by
Emmanual R. Piore, Chief Scientist of IBM, who after describing the
Project MAC system of protection of access at M.I.T., and in company
management information systems, concluded:
"...preservation of privacy rests not with machines but with
men. The effectiveness of all protective measures, however
sophisticated they may become, will still depend upon people:
operators, service personnel, supervising officers, and all
those who decide what information to put into a computer and
how to use it ... .Machines have no morals, no ethics: men have
ethics and morals. "^
The present confidentiality of individual data in Census reports,
protected by law and by administrative action, has been cited as a good
example of the protection of privacy which could equally well be applied
to the proposed National Data Center. As Kaysen has pointed out:
"The present law and practice governing the Census Bureau offer
a model for this purpose. The law provides that information
contained in an individual Census return may not be disclosed
either to the general public or to other agencies of the govern-
ment, nor may such information be used for law-enforcement,
regulatory, or tax-collection activity in respect to any individual
respondent. This statutory restriction has been effectively
enforced , and the Census Bureau has maintained for years the
confidence of respondents in its will and ability to protect
the information they give to it. The same statutory restraints
could and should be extended to the data center, and the same
results could be expected of it."-^
Carl Kaysen, "Data Banks and Dossiers," The Public Interest , Spring 1967,
(reproduced in Long subcommittee hearings, pp. 265-269).
2
Long subcommittee hearings, pp. 122-123.
Kaysen, op. cit. . (p. 268 in Long subcommittee hearings).
11-36
Despite the answers of the proponents, those who feared the
advent of an Orwellian 1984 "Big Brother" in the form of a computer whose
"button could be pushed" to print out revealing and damaging information
about an individual carried the day. The proposal for a National Data
Center is in limbo, even though there are a number of national subsystems
in different federal agencies. The benefits that might derive from
statistical analysis of aggregates based on individual economic and
social data are casualties of the fear that individual privacy might be
eroded. Even the sensible proposals of Alan Westin, reviewed earlier,
have not received serious discussion in the Congressional subcommittees
which killed the national data bank proposal.
Regional and Local Data Banks
As Westin pointed out in his Congressional testimony, there
seems to have been little concern about the privacy question in the
efforts to develop data banks in such cities as New Haven and Detroit,
or in Alameda County, California. A brief review of each of these will
indicate their purposes, recent status, and relation to the privacy issue.
1. The City of New Haven, with the help of IBM's Advanced Systems
Development Division, began in 1967 to "put the city's files on computers
to obtain a statistical profile of everyone in town.... The knowledge, for
example, that a man was crippled could be stored away for possible use by
the Fire Department. His application for welfare assistance could be
automatically checked to see if he owned a car.... The police chief
could have instant access to all the information about a suspect, (and)
11-37
the computer system could also determine if a fire broke out near a
convicted arsonist's home." This type of data bank, unlike the pro-
posed national one, does have individual dossiers, with cross referencing.
2. Detroit has been developing two data banks, one containing phys-
ical data and the other social data. "The physical data bank contains
such information as the condition of the city's residential, commercial,
and industrial housing, obsolescent structures, crime rates, and much
more" to help plan urban renewal projects. The social data bank 'Includes
statistics on crime rates, welfare, births and deaths, school truancy and
drop out rates, the occurrence of venereal diseases and tuberculosis, and
other information." Printouts of social data are made quarterly, and can
be retrieved by census tracts when needed. Apparently, social data are
used in urban planning, also, for the statement is made that "we cannot
improve a neighborhood's social vitality if we do not know what the
neighborhood's social problems are and how its people live." Also,
"in time we may even be able to devise an 'early warning system,' which
could alert us to a neighborhood drifting into instability or social
2
decline." There is no mention of the possibility that individual data
may be revealed.
"New Haven Plans a Computer Pool," New York Times , March 29, 1967.
Harold Black and Edward Shaw, "Detroit's Data Banks," Datamation ,
vol. 13, no. 3, March 1967, pp. 26-27. The authors are staff member;
of the city's Community Renewal Program.
11-38
3. The Alameda County "People Information System" includes the
City of Oakland, California and surrounding area with a population of
over 1,000,000. The system consists of two parallel but separate real-
time subsystems: (1) a central index for social services "to better
coordinate the line activities of the social service agencies like wel-
fare, hospitals, health, and probation," and (2) a police information
network. Both subsystems have data on individuals, since the first can
"identify people, locate their records, and thus respond to the thousands
of inquiries pouring in each day by telephone, by letter, and by walk-ins."
The second real-time subsystem, known as PIN, contains warrants of arrest
"to serve the 93 law enforcement agencies in the Greater San Francisco
Bay Area'.' The benefits claimed are "more efficient government" (particu-
larly in preparing reports to different State departments) and "increased
service" in better law enforcement, better future planning, and in "help-
ing welfare, health and probation social workers spot and control potential
trouble areas before—not after— the fact." The privacy question is
skirted with the note that "most information handled by counties is of a
nonconfidential nature and security requirements are not involved." But
for files that are confidential, "access is carefully controlled by per-
mitting only authorized terminals, and where necessary authorized persons —
by means of secret codes — to inquire into such files. Future plans also
call for monitoring techniques to determine what terminals are accessing
any given file."
Gordon Milliman, "Alameda County's 'People Information System',"
Datamation, vol. 13, no, 3, March 1967, pp. 28-31.
11-39
Some Implications of Centralized Data Banks
Some implications drawn from the preceding discussion may be
summarized in the following points:
1. The proposal for a National Data Center was designed primarily
to permit large-scale statistical analysis of the characteristics of
groups of individuals, households, or business establishments. The
purpose was to aid in policy planning at the national level. In
contrast, the local and regional data banks have been designed and used
primarily for administrative purposes.
2. Despite the limited objectives of the National Data Center
proposal, the issue of "privacy" of individual data or "dossiers" was
raised by Congressional committees and a number of critics of the proposal.
The objectives of the proposal were distorted by some of these critics,
who suggested that a "1984-Big Brother" type of surveillance of all that
any individual ever did in his lifetime would be available in a computer
printout "at the touch of a button." These distortions, along with some
more moderate criticism, resulted in the indefinite postponement of any
national data bank.
3. The issue of privacy of computer-based information is, nonethe-
less, a real one. A number of useful suggestions have been made to protect
access to data on individuals, and these would require both legislative
and administrative controls. Presumably, these could be made effective,
in order that the legitimate research objectives of a national data bank
could be realized.
11-40
4. In contrast, the issue of privacy has not been raised about
the (local and regional data banks, whirch^adrnTttedly do contain detailed
information about individuals. ) Sate and municipal governments have
apparently been less concerned about possible misuse of such data than
was the Congress about a national data bank.
5. Unlike the other "knowledge industries" reviewed earlier in
this paper, there do not seem to be important organizational implications
about these data banks. The proposed national center would not lead to
more centralization of government as such, but to more centralization of
final statistical collection and analysis. There is no published evi-
dence that the few local and regional centralized data banks have led
to greater centralization of governmental administration.
6. Implementation of centralized data banks has not threatened
the jobs of existing governmental personnel; nor has it drastically
changed the nature of their jobs. The main obstacle in implementing
the National Data Center has been the privacy question. Quite possibly,
greater attention to ways of protecting individual privacy in the earlier
special committee reports which recommended such a center would have fore-
stalled some of the opposition. But in retrospect it is not at all
certain that any such recommendations would have prevented the raising
of a politically-attractive and often emotional issue. Perhaps only
time and patience will overcome this obstacle.
7/2-7
'7f
i
I
i3
^•'^^th/IE
\^r
^■k ~
,
Date Due
DEC 05 75
5^R2 82#I^
,Rn7 199Z-
APR. i 3 7334
^!^^
-^^r
nov. J 3t9«r
MA^ \ 5 ''i
juNiij'ao
aPT .44 Vi
#, ..
^ * 1 ■■•ys
APR 3 79
--Si?
"^'^.7,
)ECl8l98e
Ar'""'
Lib-26-67
II HI nil iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiniii
3 TOfiO DD3 7DB 17T
3 TQfiD DD3 7DB 153
3 TOfiO D03 b71 ITE
422-
3 TOan 003 705 lEO
iiititifttiTi --'
3 T06Q 0Q3 b71 164
3 1060 003 b I'i'
3 TOaO 003 b71 17b
iiiiiiiiiliiilii ...'-
3 TDflO 003 b71 Ibfi
m<i-b^
3 TOaO 003 70E ESE
iitflfffffli"-'
3 1060 003 702 d4b
