For millennia, the healer—the shaman—has been set apart from the rest of the community as sacred, as one who trafficks in the hidden energy-forces of which his fellow men are merely awed observers or powerless victims. The nature of the work must surely have contributed to this quality of specialness: one who regularly officiates at births, agonies, possessions, fevers, and deaths is bound to come away with some unique stature in the eyes of those for whom these are once-in-a-lifetime events. But that is the least of it. Our healers have been sacred because we have needed them to be so. In the face of overwhelming impotence in dealing with the terror of illness, it has been much more comforting to seek the help of someone with one foot in the world beyond than to consult a bland rationalist, especially it neither’s medicine worked very well. The beliefs became self-proving: given the fact that by far most illnesses are self-limiting or psychogenic, the more mystery we invested in a healer, the better he made us feel. As Claude Lévi-Strauss observed about a Kwakiutl shaman, “Quesalid did not become a great shaman because he cured his patients; he cured his patients because he had become a great shaman.”
With the rise of science, rather than set itself in hopeless opposition, as did the Church, the healing profession came to embrace a marriage of rationalism and rattle-waving—the Science and Art of Medicine. Greek and Latin phrases served the dual functions of scientific precision and arcane incantation; the white coat represented the intellectual purity of laboratory research while doubling as an impressive ritual costume. “Scientific" therapies like drug prescription and injections were found to have useful magical effects as well. But marriage these days, we are told, is a dying institution. A divorce appears to be developing between shamanism and medical rationalism, with the latter refusing to support the former and being met with countercharges of mental cruelty. The patient, cast once again in the role of child, is caught in the middle.
One development that is driving a wedge between Art and Science in medicine is the growth of “information technology" —the study of how knowledge is manipulated in a given field, the anatomy of its logical structure. This new discipline examines the “decision trees" and “branching networks" that populate medicine’s epistemological forest, slashing mercilessly through the nonrational undergrowth that clutters the path. The two most important manifestations of information technology in medicine are the application of computers to diagnosis and therapy and the development of “clinical algorithms,” or step-by-step protocols, to guide paramedical workers in the management of common complaints. Underlying all is the assumption that there need be nothing arcane or extra-rational about making sick people healthy, and growing pressure, particularly with the coming of some sort of national health insurance, for more efficient, less labor-intensive means of delivering medical services to large numbers of people. Going to the doctor may never be the same again.
Part of the shaman legacy is the widespread belief (particularly common within the medical profession) that when a patient presents himself to be cured, the doctor, guided by some extra-rational power (sometimes called “clinical intuition”), forms a mental gestalt which miraculously blossoms forth as a therapeutic decision. Yet information theorists point out that the process is nothing more than a collection of many simple, smaller steps which, taken separately, are quite manageable by either mortals or machines.
Consider the way in which a doctor collects the information necessary to make a diagnosis. A good diagnostician will tailor his interview and examination from the infinite list of potential questions in order to home in, falconlike, on those areas of greatest significance to the patient’s illness. Unconsciously, he uses what information science calls a “branching program”—letting the answer to one question determine what will be asked next. Branching is a common property of many computer programs; it’s not surprising that such programs came to be written for medical problems.
Dr. Warner Slack, then at the University of Wisconsin, led the development of medical interviewing by computer. “Physicians need help with their medical histories,” he wrote in an early paper. “The family physician may see over forty patients in a day and barely has time to find out the chief complaint, let alone get a complete review of systems. And history-taking hy physicians is expensive. Talk is not cheap in medicine—it is perhaps the most expensive commodity in the clinical transaction.”
History-taking is one area in which computers have already replaced physicians to a limited extent. At several hospital outpatient departments, the once barren hour-or-two wait is now partly occupied by an interview, via teletype terminal, with a computer. in these pilot projects, the machine elicits a detailed medical history, including specific questions about the function of every organ system from genitourinary to neuropsychiatric. The laylanguage answers are translated into medicalese, findings requiring follow-up may be starred, and the completed printout is ready for the doctor when he first meets the patient.
The ability to “branch” an interview is not the only human capacity that Dr. Slack, now an associate professor of medicine at Harvard, has tried to build into his programs. The doctor can also explain questions or terms the patient may not understand; hence the computer offers four possible answers to each multiple-choice question: YES; NO; DON’T KNOW; DON’T UNDERSTAND. (In the interests of protecting the patient’s privacy. Dr. Slack has added a fifth option: “NONE OF YOUR DAMN BUSINESS,” which he has modified to “SKIP IT” for use in Boston.) Responses to the “don’t know” or “don’t understand” choices cause the machine to branch into a subprogram explaining, for instance, how to know whether you’ve ever had an electrocardiogram and what it is for, so that the interview enables the patient to acquire as well as provide information. The programs can also express sympathy for patients suffering from particular symptoms, or relief that a given organ has been producing no complaints. Finally, and perhaps most important, it is relatively easy to equip the computer with identical programs in several languages and at several levels of education—a potential blessing for inner-city outpatient departments where one bilingual translator may have to service an entire hospital having patients of many nationalities. The machine could, by introductory questions, determine the language and level most appropriate for a patient, and then turn itself on to the proper program. Of course the final printout is always in that strange amalgam of Latin, Greek, and English spoken by doctors.
Computer proponents like to point out that the traditional system does not offer clinic patients a warm, omniscient Marcus Welby. More than likely, “service” patients must wait several hours to see an inexperienced intern or resident, or a bored, resentful senior-staff doctor doing his mandatory tour of free duty. Care is often rushed, impolite, and haphazard, and may be little better for paying customers. Against this standard, they argue, a thorough, civil computer program may actually be the more humane alternative.
As for the loss of the specialness that flows from the demigod status of physicians, few tears are being shed in these quarters. “It’s time for the patient to get up off his knees,” declares Dr. Slack. “The image of the doctor as priest or magician belongs to another era. and should be left there.”
And the computerized process needn’t be dull and unfeeling, Dr. Slack argues. “A good computerized history ought to be like a good novel ... a pleasure to read page by page. Ideally, we’d have someone like John Updike writing each frame for us.” Furthermore, computerized records are legible and available wherever there’s a telephone and terminal—of great importance at a time when one patient may be seen by several doctors for a variety of complaints.
Some human doctors spend time teaching, if that word is taken loosely. So do their solidstate counterparts. In fact, during eight weeks on surgery at the Massachusetts General Hospital.
I developed a better rapport with the experimental teaching machines than I did with any of my M.D. instructors, who never found much time to devote to a mere student. Whenever I had a spare hour or two, I would sneak off to the domain of Dr. G. Otto Barnett, director of the Mass. General’s Laboratory of Computer Sciences. Going there was like entering the set of a science fiction film—in a large central cavern, the machine’s sleek guts stood in perpetual fluorescent brilliance, surrounded by transistorized attendants with names like ULTRONIC DATA PUMP. Here and there would sit stacks of fat magnetic tape disks under clear plastic domes, looking like an Oldenburg sculpture of the containers used to display chocolate layer cake on restaurant counters. The ubiquitous terminals were not the intrusive click-clack kind, but silent, high-speed models that literally spit dots of ink onto a page, using electrostatic charge to shape their trajectories into letters at a rate much faster than the eye can read. Just outside the central room was a tiny anteroom with a desk and bed: the system required a human servant around the clock.
The room I used was next door to the anteroom, a cramped area filled with teletype and videoscreen terminals. Hitting the proper sequence of keys on a console put me in touch with MUMPS, the MGH Utility Multi-Programming System. MUMPS remembered my name from previous encounters, as well as how far along in medical school I was, more than I could say for the surgeons downstairs.
Instructions were available for dealing with a variety of common problems: coma, abdominal pain, chest pain, resuscitation. The coma program was the most absorbing because it simulated actual emergency-ward situations—the patient appears with a fragmentary history from a relative (HE WAS
EATING DINNER AND GOT UP COMPLAINING OF A HEADACHESUDDENLY HIS LEGS GAVE WAY AND HE
LOST CONSCIOUSNESS) or. as often happens, no historv (THE PATIENT IS A WHITE MALE IN HLS 50s BROUGHT IN BY POLICE AFTER BEING FOUND LYING UNRESPONSIVE ON BEACON STREET). The Student is called upon to choose one ot four modes, each with its own set of choices: 1) initiate therapy; 2) take a history; 3) perform physical exam: 4) order lab tests. He can, and is expected to, go back and forth between these categories, as in real life. The computer responds as the patient would, or reports lab data on tests ordered.
An excellent teacher could equal or better a stood teaching program, but excellent teachers are appallingly hard to find, even in prestigious schools; a computer program, however, can be accessible to anyone with a terminal and phone. In simulating and teaching how to manage a cardiac arrest, computerized instruction is even more valuable than human instruction because the student can go back and try again and again, continuing to practice until he has some expertise—an opportunity not offered in real-death situations.
But what of the actual doing of medical care? It is much easier to imagine a highly rationalized approach to teaching students or collecting medical information than it is to conceive of an electronic laying-on of hands. Yet it is in the area of therapy that some see information technology as having its greatest impact. Let us consider first the easy part of healing: the straight, scientific, biologically based Science part.
The kind of complaint that might fit into the “pure logic” model ranges from the common cold to the most exotic biochemical abnormalities. During a three-month rotation at Boston’s Beth Israel Hospital, for instance, I came across a computer program that evaluates a patient’s metabolic status at least as well as the average physician could.
Diagnosis of these metabolic disorders and therapy for them have always been a challenge. The mere number of factors involved is awesome, not to mention the intricate ways that each can interact with all the others. Yet it is these very characteristics that make the problem an ideal one lor computer management. The variables are “hard” ones—they can be totally defined by numbers, unlike pain or depression or nausea. The physiology of their relationships has been fairly well worked out, and also operates “by the numbers.” Thus Dr. Howard Bleich, a Harvard nephrologist interested in computer applications to medicine, was able to devise a program, to act as a physician’s consultant, that can accept the relevant metabolic data, demand more if necessary, and, in milliseconds, provide the doctor with a list of diagnoses (in order of probability), explanation of the physiology involved, appropriate therapies, potential problems to watch out for, and even a list of recent references in case the doctor wants to learn more about the condition. The encounter takes about seven minutes and costs about $2 in computer time. And, unlike other consultants, the computer program is available virtually simultaneously at an unlimited number of locations; the Bleich metabolic program is now used in ten states, at more than forty hospitals, many of which previously had limited access to such teaching-center knowledge, and were often treating such problems less than adequately.
Dr. Bleich, who was the first to develop the computer-as-consultant idea, works in a closet-sized office in the hospital’s research building. A television terminal forms part of his desk; taped to a door is a quotation from the London Times:
That it will ever come into general use, notwithstanding value, [is] extremely doubtful, because its beneficial application requires much time and gives a good bit of trouble to the patient and to the practitioner, because its hue and character are foreign and opposed to our habits and associations. There is something even ludicrous in the picture of a grave physician proudly listening through a long tube applied to the patient’s thorax.—1834
The stethoscope analogy fits a number of Dr. Bleich’s views about where the computer will fit into the clinician’s repertoire. A “machine” to listen to the chest didn’t make the physical examiner obsolete, it enabled him to be more effective. Dr. Bleich sees coming a better matching of medical tasks to people or machines. He believes that medical problems translatable into hard facts and figures are best handled by machine. “The computer,”Dr. Bleich declares, “is immune from fatigue and carelessness; it works day and night, weekends and holidays, without coffee breaks, overtime, fringe benefits, or human courtesy. It requires occasional repair, but even in the absence of an extra machine, it is more likely than its human counterpart to be available at the moment of need. At times, the acid-base program outperforms me. After all, I wrote it during my lucid intervals, with all the relevant literature on my desk, and with more knowledgeable colleagues nearby.”
Just where programs like Dr. Bleich’s acid-base “consultant" will fit into the future pattern of medical care delivery is open to question. Will such a program remain a cute curiosity and little more? Or might it be the first of many units of computerized diagnosis and therapy that will one day handle virtually all well-defined medical problems?
The latter possibility seems more likely after a visit to the Medical Center Hospital of Vermont, where Drs. Lawrence Weed and Jan Schultz have built computer-linked records and therapy right into the fabric of patient care. Reasoning that unless the computer network is an integral part of a ward it will not be used, the Vermont doctors completely reorganized the hospital’s gynecology service, as a start, so that all patient care had to be channeled through PROMIS—the Problem-Oriented Medical Information System. All records are kept in the computer’s tape memory; when the doctor has examined a patient and wants to write a note on his findings, he must do it via a TV-type terminal at the nursing station. There’s nowhere else to write it.
Nor is this a typical passive terminal. The screen’s face is covered with twenty transparent strips, each electrically sensitive to the touch of a finger. By touching the strip over a particular phrase displayed on the videoscreen, the user can communicate complicated phrases to the computer faster than he could speak to a colleague. A keyboard also allows input of free text. If the doctor has just met for the first time with a patient complaining of vaginal bleeding, for instance, as he enters his work-up, the machine will present him with a logical, orderly series of questions and possible answers—when the bleeding started, rapid or insidious onset, and so on—branching for detail. An experienced user running his hands over the face of the terminal presents a remarkable sight, suggesting some new tactile electronic fetish. Within minutes the history, physical, or progress notes are “captured,” as the jargon has it, along with plans for therapy, in an orderly, legible format.
With the computer as the obligate center of all ward operations, all drugs must be ordered through the terminal, which will, in the process, remind the doctor what other medication the patient is taking that could interact with this one (a major danger often overlooked), what the cost will be, and what side effects to watch for. Lab tests requested will bring forth information on the results of previous tests, cost to the patient, and an estimate of how soon the newly entered test will be ready.
If this is all going on already, what might we expect in a few years? Will the computer start making house calls? Maybe.
A community that is equipped with cable TV can make any television set the receiving end of a patient-computer dialogue. The patient could dial a computerized health station on his telephone, and then, with a tone-coded dial, he could transmit a series of numbers that would instruct the machine to begin communicating with him via his own TV set. Cable TV channels not used for regular programming can be electronically split into hundreds of subsignals that, properly scrambled, could send a unique signal to any receiver in the “wired city” at the option of the central studio. There need not be any waiting time to see the electronic doctor; through the technique of “timesharing,” one computer can communicate uniquely with dozens of terminals simultaneously.
From here, the individualized program, informed by the code number, might begin by presenting the following words on the TV screen: I PRESUME
THIS IS THE JONES HOUSEHOLD. AM I CORRECT? PLEASE PRESS “1” FOR YES, “0” FOR NO. The appropriate number-buttons are on the phone, still connected with the central station. If we allow ourselves to project the development of techniques still in their rudimentary form, we might consider the case of a man who is awakened by crushing chest pain at 4 A.M. on New Year’s Day. Assuming the perfection of circuitry that could enable machines to process human speech (an ability in absolute primitivity today), the man might reach for his bedside phone, dial the appropriate number and identifying code, and say, “Doc, this is Norbert Jones,” evoking in some dark magnetic memory all the salient aspects of his medical history.
After a few questions, the program might tell Mr. Jones not to worry, but to go get the disposable electrocardiogram leads that, as a matter of public policy, would be issued to all males at their first annual physical after the age of forty. It would instruct him how and where to stick them onto his chest; how to connect the small transducer to which they are attached to his telephone to feed a constant signal of his EKG into the computer’s memory. To occupy his attention, the patient might be told that this was a routine procedure for chest pain. He would probably not be told of the cost/benefit analysis that showed the expense of the EKG-transducer system to be more than compensated for by the number of productive manyears saved when it enabled the computer to “detect potentially lethal arrhythmias and initiate defibrillation or pacemaker activity through the same wires. (The EKG monitor-pacemaker combination is not science fiction; it was developed in the 1960s by Harvard cardiologists. The problem is that many fatal arrhythmias occur within the first hour or two of the chest pain that marks a heart attack; patients therefore die before they can get to a hospital.)
Within a few minutes Mr. Jones would be in the hands of specially trained paramedical personnel who would insert two intravenous lines—one to determine his venous pressure, and one attached to a battery of waiting medicines; lidocaine, morphine, epinephrine. An automatic blood-pressure cuff would also be attached to the portable console brought by the cardiac team. As his vital signs and EKG were fed into the data bank, they would be scrutinized by logic circuits representing the collective wisdom and latest research of the world’s cardiology experts. Should a dangerous disorder of rhythm develop, signals to the cardiac ambulance would prescribe measured doses of the right drugs; should the heart stop, other drugs, and perhaps defibrillation, would be combined with the efforts of the paramedics in mouth-to-mouth resuscitation and closed-chest massage. Mr. Jones might have a good chance of living through the critical few hours that under present conditions are often the last for coronary victims who routinely die either waiting for ambulances or in vehicles or beds unequipped to handle serious illness.
By any standards, these possibilities are speculative. But they become a good deal less unbelievable when one considers that all computer technology is but twenty-odd years old; the line between fantasy and prediction becomes seductively blurred.
So far, we have considered only the strictly scientific aspects of patient care. “No fair!” cries the dwindling chorus of medical humanists. After all, they point out, many patient encounters (perhaps the majority) are motivated by completely nonphysical problems. This psychosocial dis-ease may be somaticized to gain legitimacy, but will not be well treated by a computer program oriented around blood-and-guts pathology.
What of all the lonely, confused, tired people who go to the doctor to have someone to talk to? Are they to be ignored by our unfeeling, transistorized healers? In this light, much of the present preoccupation with high technology in medicine completely misses the real issues. Moreover, traditionalists argue, a good deal of curing could be traced directly to the arcane ministrations of the doctor, regardless of their rational basis. In the response to this contention, we find a curious alignment of humanists and technocrats: If a “disease” could be overcome by belief alone, they ask, is it really in the best interests of the afflicted for the doctor to play along in this illness-affirming way? Might we instead learn to tap the “placebo power” inherent in the patient’s own mind—his ability to reperceive his “illness” right out of existencethrough nondeceptive techniques such as biofeedback, or even meditation? This would enable us to retain the positive self-cure properties of shamanism, and transform it into an honest, new therapeutic tool, dispensing with the negative qualities of the relationship. Or, we might rechannel the patient’s energies into activities directed at a more fundamental attack on the psychological or social precipitants of his complaint.
Some representatives of the left have joined the technocrats and humanists to form an unlikely trinity in defense of cybernetic medicine. They see any sort of mystification (the term was made prominent by Marx, after all) as serving to perpetuate illegitimate power relationships. A highly rationalized health care system could, they point out, end the use of health care as an opiate of the people. Medicine would no longer defuse discontent and deviance by relabeling oppression as illness, treating alienation with Valium. Further, it is argued, such a transformation could enable an underfunded system to devote all its resources to the formidable task of dealing with the neglected “real disease” of all citizens, rather than the trivial body neuroses of the middle class.
On the other hand, a quite different claim could be made by those who favor the computerization of medicine. Given the apparent need to invest our healers with supernatural auras, is it not possible that the machines themselves will take on a mana more awesome than all but the most prestigious clinicians? They will, after all, represent more collective wisdom than any single physician could hope to attain. Lévi-Strauss’s Quesalid may have been able to invoke the spirits of nature as the source of his potency, but the computers will be drawing on our own infinite wellspring of cosmic power: the holy domain of Science. The placebo effect that may reside in such a deus ex machina could put the Kwakiutl to shame.
The deshamanizing of medicine could go forward in one of several ways. The way Dr. Slack would favor could enhance the dignity of the patient by making him a well-informed full partner in the healing enterprise. Medical personnel, their time freed from routine activities by the machines, could then devote more time to the uniquely human problems associated with illness. But quite an opposite result is possible: in the name of efficiency and cost/benefit ratios, going to the doctor may be shorn of whatever personal qualities it now has. There is already a powerful bias in most medical schools away from consideration of the psychosocial aspects of disease which are so central to the patient’s experience of being ill. Machines unguided by sensitive policy decisions could turn healing into just another cybernetic travesty, in which patients’ feelings are folded, bent, and mutilated to conform to the imperatives of automation. This is the sort of outcome that is so likely that few choose to discuss it.
Will human providers of care ultimately be eliminated completely from the medical encounter? Very unlikely. Impressive as they may be, the machines are genuinely stupid in many respects, such as performing manipulative tests, or “understanding” problems not specified to the letter in their programs. Dr. Slack concedes that many medical functions will continue to be handled by people rather than machines, particularly pattern-recognition tasks and psychosocial interaction. But the future he envisions is nonetheless revolutionary. “Why must the doctor perform all these jobs? Studies have shown that someone with no sophistication in the dynamics of high blood pressure is more accurate than a doctor in making these simple measurements. The M.D. rushes through; he has all kinds of biases. I wonder whether a smart high school graduate couldn’t be trained in a few months to read a chest x-ray as well as a radiologist—or whether you could devise a training program to teach such people to interpret EKGs.” But don’t the years of medical school and internship somehow inform these interpretative decisions? Not necessarily, argues Dr. Slack. The training might act as a blinder as well as an advantage; either the shadow is on the x-ray or it isn’t; simple criteria should define whether the heart rhythm is normal or not.
Along with the waning of the physician as we know him, futurists predict the ascension of the paramedic as a key figure in the health care enterprise. In fact, for the near future, the most fundamental changes to come about in medicine through information technology may not involve anything electronic; the patient of the 1970s and 1980s may well be much more affected by the computer’s poor cousin in the information technology family, the clinical algorithm.
The clinical algorithm, or protocol, is a formalized set of instructions to guide a paramedical health worker in approaching a particular complaint. It resembles a computer program in that it is a branching network that rules in or out certain bits of data, then goes on to dictate therapy, the collection of more information, or referral to an M.D. In their training, such paramedics are familiarized with the basic physiology of special illnesses, and taught why each decision has been constructed the way it has. A well-trained paramedic would have two attributes that machines lack: a comprehension of the issues at hand (and the resultant ability to act competently in situations not covered to the letter by instructions), and the distinctively human capacities for pattern recognition, empathy, and compassion.
The natural objection: But doctors can do all this and more; why settle for less?
Protocol advocates have two answers. First, it is probably not true that doctors in practice treat these common problems better than or even as well as a paramedic with protocol. Because of its stepby-step, yes/no format, the protocol demands thoroughness, and thoroughness is more likely to be provided by a paramedic than by a busy physician who usually spends little time per case on common problems. The protocol format also allows easy audit of how well the cases have been worked up—a review that has been computerized at Dartmouth’s clinical algorithm laboratory, suggesting a new way of monitoring the quality of care, a variable well hidden from the public. These comparisons with the prevalent standards of medical care are not speculative; anyone spending time in an emergency ward or clinic will quickly see evidence of The sloppiness with which common problems like urinary or respiratory tract infections have been handled “on the outside" —antibiotics prescribed haphazardly, cultures not taken, underlying disease overlooked, therapy poorly explained to the patient, follow-up neglected.
The second half of the protocolists’ answer to “Why not doctors?” is a pragmatic one. If, they argue, there are not enough doctors to go around, and won’t be in the immediate future, doesn’t it make sense to relieve them of the time-consuming, routine aspects of medical care? Their medical energy could then be spent on those problems which require four to six years of specialized training—including care of the sore-throat or cystitis patient who, for one or another reason, has been triaged along by the paramedic.
A further objection to the protocol-paramedic approach to medical care is that it would transform healing into an impersonal, assembly-line affair, efficient but sterile. Critics see this as a step backward from the kindly neighborhood GP treating all complaints himself, no matter how trivial. But protocol-paramedic advocates argue that this is an unrealistic comparison. For better or worse, they claim, the kindly GP is going the way of his horse and buggy. The real alternative they see is a rushed, less-than-meticulous work-up by an increasingly anonymous doctor. Ironically, they feel that humanistic concerns over the quality of the healing relationship may be one of the most persuasive reasons for embracing this style of delivery. “Doctors are not chosen or trained for their ability to relate well to other people,” notes Dr. Sheldon Greenfield, an assistant professor of medicine at UCLA interested in the possibilities of protocols. “They tend to be upper-middle-class, laboratory and university oriented people, almost all male, and this often sets up walls between them and their patients. Paramedics can be drawn from the same communities as the patients they serve, and chosen for their ability to communicate as well as other skills. It’s a way of providing care for those who now don’t get any, and of de-mystifying medicine—taking health care back to the people.” A similar view is held by Dr. Slack, who sees “physicians’ assistants” playing a key role in filling in the gaps left by automated medicine. “After all, the medical profession isn’t the only source of tender loving care—many others can do a much better job of it than we do.” Yet he rejects the job title: “I’d rather see them called patients’ assistants—it is to the patient that they should feel their primary allegiance.” Again out on the fringe of the unthinkable, Dr. Stack foresees a more radical extension of the idea—write protocols for patients to use. “We ouaht to teach the patient in many cases to care for himself; he’s highly motivated to do so, and may be our biggest wasted medical resource. When necessary, of course, the protocol would let him know when to come in to see a professional.
Little by little, the specialness of the shaman is stripped away: his interviewing technique can be replaced by a branching electronic sequence, his diagnostic acumen is nothing more than a collection of thresholds and logic trees, his therapeutic decisions simply programmed probabilities. Faced with this triumph of intellectual dissection, are we to cheer or mourn?
Both. I, for one, will be glad at the passing of the omnipotent white-cloaked father figure who depended on a cowed, awed patient for his therapeutic success; who, in our day, confused and mystified the laity all too often for his own ends. Yet we need not necessarily give up other aspects of the traditional healer that are worth saving. Patients will always need a compassionate human healer to guide them through the increasingly complex network of medical resources, and this person could, if properly chosen and trained, provide the warmth and psychosocial support that machines will never be able to deliver. For otherwise, what would happen to all the interpersonal hassles, life-frustrations, and fears that people now bring to doctors— the new clergy—clothed in the acceptable garb of bodily ills?
In some ways, the question is so big it transcends information technology. We could program machines to dispense placebos and train paramedics to emulate the vapid hand-waving with which most doctors now treat such dis-ease. On the other hand, computer/paramedic-aided health care could redefine such free-floating nonorganic complaints right out of the scope of medicine: DIAGNOSTIC ANALYSIS INDICATES NO SIGNIFICANT DOC’UMENTABLE PATHOLOGY. THANK YOU. GOOD-BYE. Or. “Gee, Mrs. Davis, I’d love to do something for you, but you’re not on my protocol. ...” A more suitable solution would be for the health care delivery system to be thoroughly interwoven in a tapestry of human services ranging from legal aid to consciousness-raising groups.
Cybernetic medicine finds itself being born into a health care system dominated by two major conflicts: generalist versus specialist; rich versus poor. In some instances the machines will give an extra push to one side or the other; in most cases, they will merely be the bounty of whichever camp wins the larger battle.
It could be argued, for instance, that far from dealing the coup de gràce to general practice, information technology may resurrect it. After all, what two factors have been most responsible for the demise of the generalist? The boring, trivial nature of most complaints he sees, and his inability to keep on top of the latest diagnostic and therapeutic exotica. Give him a paramedic to handle the former, and access via an office terminal to a teaching-center computer “consultant” network for the latter, and Doc Welby may flourish in life as well as on TV.
But just as information technology could make general practice more defensible intellectually, it could also be used to fragment further the delivery of care. The American obsession with efficiency might quickly seize on the likelihood that a paramedic who does only bladder infection protocols can do them faster and with slightly more accuracy than a “general” paramedic. Physicians might be issued “passwords” to gain access only to those programs within their field of specialty. The whole enterprise might, as is now the trend, be housed in a hospital/medical school megacenter serving huge numbers of patients. Computerized, problem-oriented records could remove the last barrier to assembly-line medicine if they were used to enable any doctor to treat any patient without prior acquaintance.
The point is, we are already far along toward this Kafkaesque model even before cybernetics has had much impact. It is not the machines that are giving us such a system, it is the American public, with an enthusiastic push from the “medical-industrial complex” of hardware providers, expanding hospitals, insurance companies, and private practitioners. If we consider the present system, the machines would have the definite advantage of being able to offer higher quality care to larger numbers of people than is now available on the average. Yet the present system is not much of a standard: other nations have shown us that there are far better ways of delivering health care.
An alternative I favor would be community-controlled neighborhood health centers run as comprehensive, prepaid group practices, funded through a national health service. For the near future, physician-assistants would provide a broad range of basic services, ranging from extensive preventivemedicine programs to care of simple acute or chronic problems via protocol. Each family would be the responsibility of one small team of physicians and physician-assistants, including professionals well trained in the psychosocial “shamanistic” aspects of illness—a component of medicine we must not automate or intellectualize out of existence. Each patient would have one member of the team as his or her primary health care provider and “ombudsperson” in the journey through the medical establishment. For some patients, perhaps an elderly diabetic with heart failure, this primary provider might be the M.D.; for others it might be a nurse-clinician, or social worker, or psychologist. Computerized records would help ensure accuracy and completeness, and would enable members of the teams to cover for each other, if necessary, and yet have access to a solid data base on each patient. This could overcome the problem of slipshod “holding action” care versus overdemanding always-on-call life-styles that now mark general practice. The new technology would also enable consumers of care to exert control over the quality of the care they receive for clear-cut complaints, since protocols and computer-assisted review make possible supervision that is now beyond the ken of patients and against the desire of physicians.
Computerized diagnosis will likely be used as an adjunct to physician care for well-defined problems even in the next few years. Conceivably, in the more distant future (but probably before the turn of the century), the physician may defer in the majority of cases to the greater expertise of the collected wisdom represented in the computer’s memory.
This ideal is almost beyond belief now, not because of the technological development it assumes, which is near at hand, but because of the social development it requires, which still seems far off. In an article in the October, 1973, Atlantic, Godfrey Hodgson pointed out that the shape of the medical care delivery system has been and will likely continue to be determined by the fiscal desires of the insurance industry, hospitals, and commercial interests profiting from our present expensive healing apparatus. A locally controlled, federally funded system not based on profit is inconceivable as long as these interests predominate.
This raises the rich/poor dichotomy to which the new medical technology may have to adapt. Historical evidence suggests that unless our priorities change, cybernetic medicine, like its present-day forebear, will arrange itself along a continuum of quality: the best available care for those who can afford it, and zero, or demeaning, third-rate care for the poor. The automated expression of this national tradition isn’t hard to imagine. Outside of teaching centers providing superspecialized medicine as part of a training program, such expert care could come with a high price tag. Our free enterprise system could dictate that a program prepared by a team of elite specialists would be very costly to use, while less thorough or less reliable programs would be available for those who couldn’t afford anything better.
Two factors, technical rather than ethical, might spare us this fate. First, the potential ubiquity of the programs would rip the limited-supply factor out of the supply/demand equation that now helps to determine cost. (But bear in mind the power of patents and copyrights in preserving private interest.) Second, the probable authors of these programs are generally to be found on the faculties of medical schools. Whatever their faults, these experts are not universally greedy. Academic rather than fiscal interests could make their work widely available. Dr. Bleich, for instance, likes to compare the flow of medical knowledge with the flow of water in a plumbing system—an apt metaphor for a kidney specialist. “Hospitals don’t charge patients for how much water they use. I’d like to see computerized medical expertise made just as available, just as free. The medium could lend itself to it readily. With enough terminals and a larger reservoir of programs, the cost of using any given one could be trivial.”
Other aspects of a profit-oriented future are more discouraging. Take the uses of protocols as example. Instead of an integrated, uniformly funded neighborhood health center ideally matching tasks to doctor or paramedic or computer, with ongoing personal contact, we might well see competing moneymaking Health Maintenance Organizations that provide several plans under their proclaimed mandate of American “pluralism.” Some would offer a cafeteria-style economy model relying very heavily on protocols administered by a fleet of paramedics, even when inappropriate, while making excessive use of automated historytaking and diagnosis. Others might provide a deluxe plan of personal, specialist care, even when unnecessary. If this seems like an ungraciously cynical projection, consider our present system, in which residents, interns, or medical students provide much primary care for the poor in cramped, overbusy outpatient departments, while high-priced internists care for the wealthier. In their eagerness to extend care to the millions who now receive none, protocol advocates (as well as their computerized colleagues) will have to be careful lest they become instruments of a futuristically stratified health care business.
One final point, probably the most crucial: we must not assume, as we so often have, that any problem can be solved merely by the application of more technology and more hardware. In the case of medicine, far-reaching cultural and economic changes will have to take place before we can develop an optional health care system— changes which need have nothing whatever to do with machines or automation. A computer, or a “patient’s assistant,” can improve the quality of care or render it mediocre; it can be a means of freeing medical talent for larger questions, or just larger incomes; it can increase the dignity of healing or it can cheapen and degrade the experience. These are outcomes that are relatively independent of the technology itself; as we have learned so often and so painfully, it is the social uses to which we put these capabilities that are crucial. If we don’t allow a blind technological imperative to squeeze all that is human out of the healing process, if we don’t let lust for maximized profit margins contaminate even more of medicine, these tools may play a role in ending the crisis of health care delivery we now face. But if we choose to approach these problems as we have approached so many others in this century, even pulling out all the plugs won’t help.