“Where will we be in ‘93?” is the title that I have been given by the organizers of this meeting. I am pleased to have a shot at it even though I would probably not have been egotistical enough to propose such a broad range talk. It requires not only research and ability to extrapolate existing state of the art, but it also forces decisions with respect to the probable way our medical hardware and software will go in response to likely major reconfigurations of the medical systems we serve, under social, political and economic pressure in addition to normally expected technological advance. Several of the feasible paths appear to be mutually exclusive.
We are under fierce international pressure to catch up with and preferably to step ahead of Oriental and European automated manufacture and the associated medical robotic technological development that the Japanese at least, consider a closely linked research and development. They have demonstrated their ability to make such applications to new fields quickly and economically. We had better learn these managerial and manufacturing skills.
We must continue to serve and update efficient operation of current medical practices, that are excellent extensions of the innovations of twenty years ago, remembering that in practice one can expect to rediscover and exploit at a higher utility level, any major new idea about once per human generation.
Medical science now has two new major acceptance areas that have, in effect, just opened up. These must go into our plan. The unexpectedly rapid and high level of success with image synthesis in NMR scanning, added to the intensive and successful hard slugging to bring CAT scanning, utilizing X-ray and ultrasonic modalities, into everyday acceptance with expectation of diagnostically useful results, means that we can expect to get other image enhancement and feature extraction procedures requiring heavy computational burdens to be well received. Clinicians are becoming more receptive to computational imaging aids as they learn that they do not have to understand the hardware and software personally.
At our meeting last year at Santa Barbara, I alluded briefly to the emergence of substantial fuzzy clusters of really quite different kinds of technically implemented health care that we must be prepared to deliver in response to patient pressures, as patients learn that medical attention can come, not only in the hospital or doctor's office, but also in the home and in utility “supermarket” economy forms that can still be good medical care and a real bargain.
For the individual (or family) who is really prepared to participate and pay for it, there can be, even with currently available technology, superb quality, very highly personalized care with home computer-related instrumentation.
For the traditional patient there must still be the classical service in which
the patient has complete confidence in the doctor and his recoiwnendations and
must be cajoled and scolded into reasonable compliance.
Quite suddenly we are under social economic pressure to improve the cost effectiveness ratio of health care, in the face of ever growing precedent for retroactive liability expenses that may make us reluctant to introduce new instrumental diagnostic and therapeutic devices. We must look for an intensive confrontation with forensic and regulatory crises as we begin to implement more sensing, actuation and closed loop hardware with built-in elaborate decision-making software. We will have to settle the question of when and under what circumstances robotic medical equipment is either licensed to practice medicine, or can be authorized to deputize for the physician as does the nurse or medical technician.
I would like to get down to cases a little more specifically. Because all of us here share a special interest in cardio logical biomedical engineering, I can take an illustrative example from our own field that typifies the broader problem of medicine in the Nineties.
Just about two decades ago I found myself offering consultative advice to three of our pioneers in bringing computer-aided electrocardiographic analysis into operation. The three were Cesar Caceres, Hubert Pipberger and Max Woodbury. Each of them damned me rather soundly for listening seriously to the stupidities of the other two, and almost suggested that I was consorting with the enemy. Yet, each had a different and sound philosophical basis for his ideas. We have still not achieved unification of these approaches into one best path and may want to preserve more than one alternative, but near optimal combinations may be only two or three years away.
Cesar said point blank “I will emulate the good electro cardiologist in detail and in his language, appropriate or not. He will understand my machine results and empathize with them and use my programs. We can reeducate him and the machinery when we have time and spare profit to put back into research.”
Hubert stubbornly insisted on working with state of the art vector-cardio graphic theory while still insisting on, and emphasizing his clinical intuition and experience with patients, as important guiding factors. Again, he was right, and contributed much to the genesis of the many present—day programs that are becoming almost what publishers call “boilerplate”.
Max wanted to introduce control theory and sophisticated statistical analysis
from the start, when there was not really enough know-how to do the first two
levels. The time was not yet ripe.
I do not think that Woodbury really had homeodynamic analysis in mind at that time, but he sharpened some of the tools that we should now incorporate incur advanced multivariate analysis and, let me add, our structuring of epidemiological data banks.
Put these all together and you have guidelines that will last into the Nineties. Develop and improve reliable “boilerplate” programs to serve busy, non-research, physicians well and without mental effort, in familiar terms. Beside efficient, economical day-by-day diagnosis, good pacemakers and monitors and even arrhythmia analyzers can be created quickly by this approach.
The approach using current theoretical state of the art medically can, with a little delay, feed into programs that are parametrically individualized to fit the patient, incorporating any of the newer features. We can convert a bit of the electrocardiographic noise back into signal information if we will judiciously use phase-lock-loop techniques, individually normalized stress response, and in particular, multimodal, time-coherent records from diverse transducers. At present we get almost no vascular dynamic or perfusion impedance measures out of electro cardiology, nor do we even enter into our records and programs, the fact that a patient is substantially chronobiologically time-dependent and does not respond like a nice linear differential equation to partial derivative analysis.
Here is the third area where real advances in medicine can be made by virtue of computers, and the ability that they have to lead us into new and really quite different algorithmic figures of thought.
It becomes ever more evident that much disease is not due simply to invasion by microorganirns, malnutrition or inevitable failure by aging, but is actively “programmed” into our control system. We “learn to be ill”, not only psychosomatically but physiologically. We learn each month of hitherto unexpected CNS biochemical vectors and modes of record keeping and control that were not seriously suspected a decade ago, despite seemingly inexplicable onset of disease, or unexpected recovery.
We have to confront this dynamic system head-on in multivariate, time dependent form even though this is really a fighting challenge that most of us will ignore, hoping it will go away.
On quite a different tack, we must listen for valuable filtrate in the lore
of ancient, primitive and only partially credited medicine where we may find
gems adaptable to modern medicine without accepting the associated rituals and
cryptic jargon. How does TENS stimulation for relief of pain really work? At
what age do we program our individual hierarchical systems of cardiovascular
and other physiological control and supervisory skills? Are these responsive
to biofeedback or other methods of reprogramming? What is the residuum to be
filtered out of acupuncture, meditation and the many other folk derived, pseudosciences
that have never been formulated algorithmically?
While I am willing to talk at length philosophically at some other time, I am reminding you here, that we are technically prepared, with computer support, to make multivariate transponsive medical diagnostic measurements, but don't do so because it is not yet recognized as completely
feasible and a wave of the future. Let us put perhaps ten percent of our effort into this new frontier. Let us make homeodynamic medical practice legitimate and active.
Allow me to examine the portents and consider where we may be in ten years. If we stimulate patient participating, computer assisted, health care as a natural sequitor to the huge movement of microcomputers into the home, the office, and into the tools of industry, we can envision the home medical computer plus modular interface support at perhaps 500 - 1000 1983 dollars per home, about comparable to the television-video-stereo complement. The aggregate is initially frightening and then stimulating as a source of valuable entrepreneurial development of something like fifty American billion dollars. Even if this level is not reached immediately, but some is added for the hospital and clinical as well as for the health record keeping, and then a share for the communication and the diagnostic machinery, we are looking at a “Silicon Valley” size project, especially if we include the large market for biomedical computer aids for the handicapped, a topic on which I would like to say more later.
Perhaps it is time to recapitulate this interesting, but systematically autonomous, set of proposals for our envisioned status in ninety three.
Obviously, the most immediate attention, and perhaps a majority of our current effort, has to go toward steady advance of our now “classical” computer-aided Cardiology. This would include smarter, but less expensive, diagnostic “carts” with autonomous, dedicated, tractable, stand-alone computers, versatile computerized monitoring and control systems, implantable, portable or bedside, singly or in combinations, as well as multitransducer.
Biological scientists, and sometimes physical scientists as well, are accused of “anthropomorphizing” their theory and hardware; that is to say, treating it as if it had human attributes and almost verged on the animate. I am sure that I have at times felt that one of my electronic computer breadboards was surely possessed by a demon of some sort and very often such hardware does respond favorably to a small spanking.
I would like to urge us, in developing the first area, to adopt a return to more anthropomorphic interpretation. In our meticulous time- data-series analysis we have tended to personalize the several P, Q, R, S, T and other potential vs. time wave forms as if they were real biomedical entities. I believe that, to an appreciable extent, this imaging has inhibited our thinking in terms of physiological, emotional and disease entity universals.
Some ten years ago we were practically forced, by lack of appropriate computer
hardware and modeling software, into using lumped dipole models and mathematically
dubious wave-front sheets of myocardial excitation. Now we can indulge in interpenetrating
domain models of cardiac excitation and recovery, or repolarization, that are
much closer to electrophysiological honesty without being monstrous consumers
of computer number crunching. With our superior computer graphics we can now
begin to display, in real or slow motion time and color-phase-space within the
dynamic geometric heart space, the events of the sequential normal or abnormal
beats for much more intuitive understanding and comparison with “normals”.
This first line of development does not require extraordinary facilities, financing or major system redesign. Obviously it demands hard work and some effort toward basic standardization of measurement, displays, data storage and retrieval, and the accompanying new figures of thought, but it can be evolutionary - there is no need for revolutionary rearrangement.
While cardiovascular and cardiopulmonary diseases, their appropriate preventive maintenance, diagnostic analysis, monitoring and therapeutic care constitute a substantial part of the proposed programs two and three, they are only one segment of the medical areas that must more or less even-handedly be integrated into our intertwined health care system, in which we strive, over the many medical disciplines involved, for compatibility combined with flexibility to meet the diverse needs or desires of the individual client or family seeking optimal medical care in personally satisfactory form.
We all know that the relatively competent and well intentioned HMO services, to which most of us subscribe in one form or another, usually provide negligible preventive maintenance beyond casual advice against overeating or smoking and perhaps routine supply of antihypertensive medication or insulin for diabetics. One has to be prepared technically, and almost offensively insistent, to get these busy systems to provide “how-goes-it” tests and the necessary accompanying careful longitudinal study of the individual medical record to formulate true preventive medical maintenance. They do provide access to usually good emergency and acute care as for surgery or intensive hospital care.
You have probably all met the GOOMER (Get Out Of My Emergency Room) who tries to free-load on this emergency service, usually available without appointment and often at low or zero cost, to replace ordinary individual physician or HMO service. He is one of the products of this squeeze between rising cost of conventional medical care and dissatisfaction with semi-automated HMO service.
The medical supermarket service caters to this kind of walk-in client and is
likely to expand considerably beyond the present scattering of medical emergency
shops, dental, ophthalmologic, physical conditioning and health checkup offices
that are springing up in many shopping centers. Most of these facilities can
operate without very expensive or highly sophisticated instrumentation beyond
ordinary X-ray, electrographic, microscopic or biochemical analytic systems
and the like. True, there are beginning to be “do-it-yourself” packages for
heart rate, blood pressure, biofeedback conditioning, etc., that may need customer
tutorials or well—trained attendants.
Our second channel, the modular computerized health-care-in-the-home technology, is the one where enormous expansion can be expected within the decade we are examining. A well debugged ECG-VCG program with arrhythmia analysis and monitoring capabilities need not be particularly more elaborate than a WordStar or high class Star Wars program, and there is a perfectly delightful variety of medical measurement, treatment and rehabilitation games we can play at home with modules attached to our basic black medical box. This, of course, guarantees the emergence of a group of highly sophisticated specialized hypochondriacs, but it also allows very comprehensive, inexpensive health care centered about the special needs of individuals, and not including unneeded modules. It allows individual selection of the bare bones austerity package or the luxury system with all the bells and whistles.
For this trick we really need to provide regulatory procedures that are flexible, safe, yet permissive, with full rationale statements and interpretations of cost benefit ratios in dimensioned and scaled form.
If you want to be amazed, ask an audience of perhaps fifty individuals of some medical, technical or layman group, after explaining the concept of computer assisted family participating health care in the home, what they would especially like to have as medical modules in their own homes. Each will easily identify 20 to 50 modules, many of them quite reasonable and very different according to the interests and knowledge of the group. You will discover wanted medical measures and services that will surprise you with their strangeness and reasonableness. I have a collection of over a hundred) and only a few of them are frivolous or stupid.
To meet this market, there must be planned modular compatibility even between competitive suppliers. There must be system planning to allow home modules and their users to seek advice from central computer information files and processors. Intensive tutorial backup must be provided in other than textbook form. This challenge I find frightening, with its complexity and difficulty of managing, marketing and legalizing, but still delightfully challenging.
We dare not attack this modular home health care design at a sub critical level
of effort and funding or it will fail, and it will require the very best of
industrial, academic, governmental cooperation and a great many of the new style
competitor-cooperative programs that are now being explored, in which many members
of small, medium and large industry each take part in a systems-coordinated
many parts modular development.
The fourth and final track for our decade of development makes for strange bedfellows. If we are to see health care decentralized and distributed into several major clusters, geographically and by patterns of operation, e.g. home care, HMO care, supermarket care or even automated private physician care, we need a new or refurbished telecommunication, medical data processing, information storage and retrieval, and transponsive tutorial facility as pervasive as our telephone system, and deliberately economical and adaptable by design.
The medical data network represents a large initial expenditure and a huge effort in reeducation, but promises perhaps the cutting in half of our current medical-health maintenance bill, with opportunities to improve rapidly with advancing technology.
Immersed in this tutorial concept is the essence of the practice we will all have to adopt soon in repairing our new and sophisticated machinery, our robotic gear, medical and otherwise, and in filling in gaps in our own knowledge and practical skills.
We are still issuing manuals several hundred pages thick with our major medical
instrumentation packages. See by a first hand test or two whether your own good
medical electronic technicians and their supervisors can and do really take
the time needed to learn these elaborate procedures, by detailed reading and
testing, to the point where they can actually service, maintain, repair and
certify a chosen package. They need to have the material supplied to them by
computer graphic, alphanumeric, and indeed verbal advisories. These advisory
services can draw on central data bases at some distance geographically without
much increase of cost. There should be much less calling for the expert serviceman
while the system remains dead. Except for some special attention to “nominal”
hardening against possible deliberate enemy action, we could today begin on
this kind of service network for both the professional and the do-it-yourselfer.
Perhaps I should remind you that we have a growing and quite progressive body
of AVTI institutions that can provide personnel and middle level management
in this context with many specially trained technicians in prosthetic, rehabilitative,
medical electronic and other appropriate disciplines. These individuals often
have hands-on training that professionally trained engineers and bioengineers
lack, even at the graduate Levels.
By a most fortunate series of coincidences, I find myself a member of a commission that our Minnesota state governor has set up to establish a Minnesota World Trade Center. This trade center is viewed as a “one stop shopping center” to market Minnesota products within this country and abroad, and to regain a more favorable financial and labor balance, by extending the basically strong agricultural, mining and electronics-computer industries that have been state mainstays in past years, despite what many businessmen consider unfavorable tax and cost-of-doing-business climate.
Because we do have, in Minnesota and its adjoining Midwest territory, unquestionably excellent resources in industrial electronic computer development, manufacture, and support, as well as fully justified reputations for medical know-how and facilities, I am intrigued with the possibility of generating a Midwest Computer-Aided Medicine Valley among our “10,000 Lakes”. This effort need not be confined to Minnesota, but could gather strength by including our neighboring centers of medical, computer and robotic talents.
To the best of my knowledge, there is at this time no consortium, governmental agency, academic institute or industrial organization with breadth of scope and competence, sufficient motivation, authority and financial as well as political stature to guide and stimulate the computer assisted health care that is proposed, on the four tracks that have been outlined. Perhaps it is for lack of such an action organization that so little progress has been made to date.
Without being an especially good marketer, politician, financier or industrial manager, one can draw up the basic specifications of an organization, perhaps unique and self sufficient, but more likely one of a group of centers at diverse geographical locations that could make the desired progress a reality.
Drawing on the experience gained in other countries, especially in Japan and western Europe, and the needs expressed here, one feature of such an organization emerges strongly. The coordinating organization must have free access to academic, governmental, commercial and industrial staff, facilities, management and finance, but emphatically should not be dominated or “owned” by any one of these interests. This has emerged as a common feature in related foreign programs of stimulated research and development.
In drawing up a design of a “Center for Innovation and Technology Utilization”
some months ago, I listed fifteen functions that should be incorporated in,
or served by, a center of this type. As a Center, or group of Centers of the
type proposed in this compendium might serve our computer-aided medical systems
needs, I have appended that list and invite your discussion, advice and additions
to or deletions from, the list.