CHAPTER 7

Matters of Life and Death
JN PRACTICALLY NO OTHER FIELD ARE PREDICTIONS AS -L frequent as they are in medicine. While even meteorologists are not normally expected to publish more than one daily bulletin, doctors must prognosticate the whole day long. The fact that they do not have to make their pronouncements public, does nothing to diminish the gravity of their responsibility, since the patient whom a doctor tells whether he will recover quickly or whether he will have to count on a long spell of suffering, listens to every one of his words and to every shade of meaning as if the doctor were an oracle.
Moreover, when it comes to an operation, not only the patient but his dependents also, become intimately concerned in the problem. It is before the dependents that the doctor must weigh up the risks involved in an operation or its omission, and if he makes unequivocal pronouncements he runs the risk of gaining the reputation of recklessness. If, on the other hand, he is careful to put forward all the pros and cons of the case, he may undermine his own authority, and thus the patient's faith in him and his cure. Medical prognosis is therefore very much like diplomacy — the doctor cannot simply discuss the objec- tive facts, but must dress them up in such a way as to make them palatable to his audience.
Nor does that exhaust the range of his prognostic activities. National or private insurance authorities must be told how long a patient is likely to be incapacitated. Life Insurance Companies demand clean bills of health which involve long range prognoses, immigration officials may make similar demands, and in many countries pre-marital health certificates are obligatory. All these, though based on purely diagnostic investigations, are in fact implicit prognoses, for otherwise they would be a waste of time.
198
MATTERS OF LIFE AND DEATH 199
Hippocrates' teaching
One might have assumed that so wide and important a branch of medicine would play a correspondingly large role in the education of medical practitioners and in medical literature. In fact, while diagnosis and pathology represent large and inde- pendent branches to which countless books and investigations are devoted, prognosis is a mere step-child, and is practically ignored by teachers and students alike. Thus an investigation of 100 English, French and German text books on general medicine and on the most important special branches of medicine, showed that less than 1 % of the subject matter was devoted to prognosis. In the great majority of these books, the words "prognostic" or "prognosis" were found neither in the table of contents nor in the index. Nor has much special literature been published on the subject, so much so that, in Germany and Austria for instance, no more than three compi ehensive works on prognosis have appeared during the last 50 years. ^ Actually, things were not always like that, and until quite recently, prognosis was much more greatly emphasised in medical literature. This was in the great tradition of Hippocrates, who founded scientific medicine almost 2500 years ago. Hippocrates himself wrote two books on the subject of prognosis and enjoined his students to pay very particular attention to them since, without prognosis, they could not hope to gain their patients' confidence. True most of Hippocrates' prognostic ideas are of only historical interest today, for he concentrated chiefly on facial expression, posture and movements. Thus sleeping with an open mouth and outspread legs or gesticulation in sleep were unfavourable signs, and the Facies Hippocratica — a prominent nose, sunken eyes, sunken temples, tense skin and yellow or black complexion — was a criterion of impending death. However, Hippocrates did not restrict his prognostic comments to such external factors alone. One of his favourite methods (which caused more con- fusion than enlightenment) was to make prognoses from the appearance of the patient's urine. According to his pupil Theo-
1 T. Brugsch: Allgemeine Prognostik (Berlin-Vienna 1918), H. Curschmann: Lehrbuch der speziellen Prognostik innerer Krankheiten (Stuttgart 1942) and H. Winter; Die Individualprognose in der inneren Medizin (Vienna 1950).
200 CHAPTER 7
philus, whose dictum on this subject became the dogma of many generations, cloudy and reddish urine on the fourth day of a fever was a clear sign that the crisis would set in on the seventh day.^ Particularly under the influence of Arab physicians, the inspec- tion of urine became a medieval criterion for distinguishing "serious" medicine from quackery and magic.^ From urine inspections, not only the shrewdest diagnosis but also the most far-reaching prognoses could be made. Characteristically, the moment that serious science discarded this questionable method, charlatans took it over to inflict it on mankind to this day.
It is quite understandable that the many errors and fallacies of past methods have made medical practitioners sceptical of prognoses as such. Unfortunately, this scepticism is misapplied, for doctors simply cannot help prognosticating in practice and, short of learning this art from reputable sources, they must rely on their own intuition and on their own limited experience. There is little to guide them in the literature, and what statistical data there are^ — except in surgery which occupies a special position — are still so utterly inadequate and contra- dictory as to be merely confusing. For this reason many practitioners — and the leading and most experienced doctors among them — make a virtue of necessity by maintaining that prognoses differ from case to case. And so they do, but if we were to apply that criterion to diagnosis and to therapy, there would be no medical science based on scientific principles at all.
In fact the difficulties of modern medical prognosis are largely social. With the decline of the family doctor, the number of medical practitioners who observe their patients from birth to death is growing smaller from year to year, and hospitals and casualty wards have such a great turnover of patients that they can do little to observe the development of a given ailment. The specialists, who, after all, write most of the literature, are even less likely to be able to carry out long term observa- tions, since the moment their patients improve, specialist services are generally dispensed with. This lack of continuous contact between doctors and patients makes it extremely difficult to develop a science of prognosis — medical congresses and exchanges of views notwithstanding.
^ M. Weiss: Diagnose und Prognose aits dem Horn (Ulm 1954) p. 19. 2 Douglas Guthrie: A History of Medicine (London 1945) p. 90.,
MATTERS OF LIFE AND DEATH 201
1001 diseases
All medical prognosis must consider three main factors: the aetiology of the disease, individual characteristics of the patient — constitution, hereditary defects, age, sex, habits, and social position — and the possible means of therapy. The most basic question is that of the aetiology, i.e. whether the causes (and the development) of a given disease are so general as to allow predictions from past experience. If that is the case, the prog- nostic problem is as good as solved in the majority of cases. Unfortunately, things are not as straightforward as laymen commonly believe. It is a fallacy to think that once a doctor has made his diagnosis, the rest follows automatically — the typical course of many diseases is largely unknown. There are so many possible developments that not even the most experienced doctors can speak of a norm. At best, they can distinguish a a number of typical, though uncertain, pathological expressions of the same illness. Moreover, the number of diseases has increased tremendously, at least in name. There are now over 1000 medical conditions, endemic illnesses excluded. In estab- lishing a "new" condition, its discoverer need not specify its exact aetiology ; all he need do is to show that the acute phases are clearly distinguished from those of any other illness. The prognosis in each and every one of them must therefore be established after a great deal of laborious research, which usually takes a few decades, and often tells us no more than that one disease may have a number of different results.
In contradistinction to other branches of science, and modern physics in particular, where recent theoretical revolutions have led to a general drive towards unification, medicine is under- going a period of decentralisation, i.e. of compartmentalisation. Undoubtedly, this decentralisation has many advantages, and no one in his right senses would desire medicine to revert to the rigid dogmatism from which it suffered until the 17th century, and which threatened to make its come-back at the beginning of the bacteriological era. Even so, so unexplored a field as medical prognosis could only benefit from a set of guiding principles.
Luckily, prognosis can never be completely wild and fancy- free, for it is bound to a factor independent of all changes of
202 CHAPTER 7
theory and practice: time. To prognosticate means to predict in time rather than to prophesy about the dim distant future. Thus medical prognosis is unconcerned with the average age of man in the year 3000, or with what fatal diseases will arise at that time. Its proper field is the prevention and cure of diseases, here and now. True, medicine can tell us much that might conceivably affect the health of future generations, but it does so only by the way.
For that very reason, medical prognoses are so often terribly pessimistic. Thus a wave of gloomy predictions was made some 20 years ago on the course of infective diseases, only to be upset by the discovery of sulphanilomides and antibiotics. It might be argued that, medical knowledge being what it was, these dis- coveries could not possibly have been foreseen by anyone, and that the predictions were therefore quite logical at the time. However, this is a form of logic that does not strike us as particularly revealing.
All the same, not even modem doctors can act any differently, for they, too, must make static predictions, i.e. assume that there will be no radical changes in therapy within the next thirty years. Dynamic forecasts so characteristic of those economists who speculate about economic changes in a.d. 2000, are quite outside the scope of medical prognosis, since the sensitive objects of its enquiries would resent mistakes most emphatically.
Infections
Even so, doctors can afford to be somewhat more daring in some of their forecasts than meteorologists, for unlike the latter, they can make accurate or highly probable predictions about a host of phenomena. One striking exception is the case of coronary thrombosis, for although some of its predisposing factors are known, it is still impossible to predict whom it will afflict fatally. All we know is that 40% of all cases die within the course of a few days, and that people with distended hearts, severe diabetes, syphilitic aneurysms, and certain pulse anomalies are unlikely to recover.
But apart from severe heart attacks, doctors nowadays con-
MATTERS OF LIFE AND DEATH 203
front acute diseases with far greater certainty than ever before. Not only can they intervene in time, but they can also predict expected changes in the overall picture of the disease from day to day, and know when the crisis and final recovery will set in. Still, the possibilities of correct prognosis differ from illness to illness.
Simplest of all, are prognoses of infections, since the germs responsible usually develop with clockwork precision. In most infections it is therefore relatively easy to predict the interval between the actual infection and the appearance of the first symptoms: 2 days in diphtheria, 2-4 days in scarlet fever, 14 days in measles, 10-21 days in typhus, and about 3 weeks in syphilis. Admittedly this knowledge is rarely of practical value to the patient, since few people can tell that they have been infected and thus rarely consult their doctors before the onset of the symptoms. However, once an infection has been diagnosed, its course can be predicted much more accurately than that of most non-infectious diseases.
Some infections follow a fixed rhythm, and malaria sufferers know precisely when to take quinine against the next bout of fever. Still, only a doctor can tell them whether their malaria is due to Plasmodium vivax or the much more dangerous Plasmodium falciparum, and predict the future course of the disease accord- ingly.^ In the prognosis of pneumonia it is also a matter of importance whether the disease has been caused by pneumococci or by far more dangerous streptococci. Moreover there are distinct types of pneumococci of which some are more virulent than others.
In these, and in a number of other diseases, the microscopic picture is therefore an excellent prognostic criterion. In other cases, chemical investigations replace the microscope, but in neither case is medical prognosis advanced enough to rely on such objective tests alone. Laboratory investigations must always go hand in hand with direct observations of the patient which, in most cases, are still based on methods used in Hippocrates' time.
In one of the most recent works on prognosis, for instance, the section on pneumonia contains the following "hints" about the
1 C. F. Craig and E. C. Faust: Clinical Parasitology (Philadelphia 1945), p. 21.
204 CHAPTER 7
patient's appearance and behaviour: "Subicteric (i.e. slightly jaundiced) colour is a sign that the disease will run a violent course, but, like the delirious state which accompanies lobular pneumonia, its prognostic value in individual cases is uncertain. Deliria accompanying lobar as distinct from lobular pneumonia, on the other hand, are usually certain signs that the outlook is poor. Similarly, meteoristic distensions of the abdomen accom- panying pneumonia, must be evaluated very carefully." Side by side with these pronouncements, the author — who, by the way, is a leading Viennese physician — also gives prognostic advice that might well have stemmed from an Aesculapian priest in ancient Epidaurus: "If a case of pneumonia is unusually bothered by flies in a well-aired ward, the prognosis is bound to be bad. The same is true also of a number of other diseases, and particu- larly of jaundice. While I would hesitate to explain this fact, it is particularly striking when a number of patients suffering from the same disease share a ward, and when some, and not others, are bothered by the flies. ^
Modern difficulties
If medical prognosis seems more speculative today than it was fifty or a hundred years ago, this is due to the paradoxical result of recent medical advances. Once a typical disease has been diagnosed, it is, of course, fairly easy to predict the subsequent stages, but unfortunately non-infectious diseases are rarely typical. While they are much better understood than they were, say, 100 years ago, the new knowledge consists largely of details that do not greatly affect prognosis, or of the introduction of countless complications where no complications were previously suspected. Hence, many doctors can no longer see the wood for the trees, and while it seems likely that our present analytical stage will once again give way to a new synthesis, that day seems very far off.
Medical advances have had yet other repercussions on the art of prognosis. During the past 100 years medical science has managed to eradicate some of the most murderous diseases 1 H. Winter: op. cit. pp. 61-62.
MATTERS OF LIFE AND DEATH 205
almost completely, and to reduce the severity of a host of others. During the second half of the last century acute infections were being vanquished by such sanitary measures as running water and public drainage. Cholera and typhus were overcome in that way, while compulsory vaccination put an end to smallpox. Finally, puerperal fever was banished from hospitals and homes by antiseptic measures. During the first half of our century, there began the great battle against the chronic infections: syphilis and tuberculosis, and here, too, the successes were extra- ordinary.
If we classify diseases according to their duration, we may say that the last fifty years of the 19th century vanquished mainly infections of short duration and with a very high mortality rate, while the first fifty years of the 20th century vanquished more chronic infections. In either case the battle was against infections whose course could be prognosticated fairly easily, and not against the most frequent causes of death: non-infectious diseases which have far more uncertain prognoses — ^particularly diseases of the heart. In other words, diseases with simple prognoses have become relatively rare, while chronic diseases with uncertain prognoses have become more common, if only because we have learnt how to keep chronic patients alive. For instance, pernicious anaemia which used to be fatal, is fatal no longer. Now, the longer the duration of a disease, the greater the number of possible complications, and the greater the role of such unpredictable individual characteristics as specific resis- tance.
This is the reason why physicians, no matter how often they are called upon to make prognoses, usually avoid publishing their prognostic opinions. One of England's leading heart specialists Sir James Mackenzie (1853-1925) used to say that "no doctor lives long enough to write a reliable book on prognosis",^ and modem doctors would refrain from writing such a book even if they lived to grow as old as Methusaleh. Thus all those medical textbooks which devote a special section to the prognosis of, say, heart diseases, usually preface their remarks with the comment that no fixed rules of prognosis exist. ^
1 D. Guthrie, op. cit., p. 56.
2 R. P. McCombs: Internal Medicine in General Practice (Philadelphia and London 1947) p. 94.
206 CHAPTER 7
Things are a little better with psychiatric prognoses. Most severe mental disorders are chronic, and psychoses, in particular, persist for a very long time — electric shock therapy and the more recent chemo-therapy notwithstanding. Quick cures are very rare indeed, and psycho-analysts, for instance, make a point of warning their patients against them. Permanent cures take years rather than months,^ and psychiatric prognoses are very long- range predictions, not only about the therapy itself but also about possible relapses after the treatment has been terminated. Relapses are, in fact, the bugbears of psychiatry, and patients can only be pronounced really cured years after the disappearance of all symptoms. But while psychiatric prognoses are apparently the most difficult to make, psychiatrists have the advantage over their other medical colleagues that they can generally follow up their patients' progress. Schizophrenia and manic-depressive states, for instance, are described in thousands of detailed case histories, giving accurate details of relapses, including those which occurred 19 years after termination of treatment.'' It would appear that though depressive states and schizophrenia — a disorder that is particularly widespread in the U.S.A. — can be cured more readily with recent techniques, the number of relapses, too, have greatly increased. Thus, while 72% of all manic depressives treated, for instance, in the Pennsylvania Hospital in Philadelphia could be discharged as cured, almost 10% of the "cures" had relapses during the subsequent five years — and five years is a very short time when it comes to manic depressives.
Although such statistical data cannot be applied to individual patients, they are nevertheless of great value to doctors who have to decide when, and under what conditions, mental patients whose condition has perceptibly improved, may be released.
Surgery = Foresight
A quite different prognostic situation exists in surgery. No
1 L. S. Kubie: Psychoanalyse ohne Geheimnis (Hamburg 1956), p. 4S.
2 The Tear Book oj 'Neurology, Psychiatry and Neurosurgery 1945-1955 (Chicago 1955), pp. 285-288.
MATTERS OF LIFE AND DEATH 207
surgeon worth his name will perform a major operation without weighing up all the risks involved. Now, unlike physicians who seem to have a horror of applying statistical considerations to individual patients, surgeons have long ago realised that only by classifying certain types of operation can they make any surgical decisions at all.
Strictly speaking, surgical prognosis is therefore based on past therapeutic results. True, other branches of medicine also consider therapeutic results, for instance, when using new techniques, but once the technique has been applied, statistics are ignored. Few medical practitioners will be able to tell you in exactly how many cases a new method has worked and in how many cases it has failed. Surgeons on the other hand, keep a strict record of even the simplest routine operations, with the result that they have accumulated a wealth of up-to-date data, which are quite unlike the collection of outdated observations on which most physicians have to rely.
Above all, surgical data contain valuable hints about the prognosis of diseases. Operations are, after all, desperate measures that are only taken when everything else has failed. In many cases, surgical interventions decide the future course of diseases, and even where they do not produce permanent cures, they, at least, get to the bottom of the trouble and lead to better prognoses. However, all this is only a minor aspect of prognostic surgery.
Even before the surgeon is in a position to predict the future course of a disease on the basis of the operation he has per- formed, he must make two preliminary prognoses. First of all, he must tell the patient how long he must expect to remain bed- ridden if all goes well, and how long he will have to convalesce. Precisely in the the case of relatively minor operations, this part of the prognosis is most important, since frequently it decides the patient one way or another. Thus if the con- scientious surgeon tells a patient that he may not be able to go back to work for months, the patient may not be able to spare the time. This type of prognosis is therefore a very delicate matter, not so much for medical as for psychological reasons. Purely technically speaking, however, it is so highly developed today, that surgeons rarely make any mistakes.
The second and more crucial part of surgical prognoses deals
208 CHAPTER 7
with the operational risks. Here statistics are most important, for although quite novel operations are frequently performed — or else surgery would never advance at all — 99% of all opera- tions carried out in hospitals throughout the world are based on time-honoured methods. Thus the patient who confides his life to the surgeon's knife is anything but a guinea pig.
However, a surgeon's decision to operate is not based on statistical considerations alone but also on the evaluation of the risks the patient runs if the operation is not performed. In other words, he must balance two independent prognoses: the patient's individual condition and the general surgical risk. In acute appendicitis, for instance, delay may be fatal while the operation itself is relatively simple, whereas in many heart diseases the opposite is often the case.
If the pros and cons are fairly equal, few patients will agree to submit to preventive operations which can be postponed and, for instance, in some cases of cancer in which the dangers of delay and the surgical risks are so nicely balanced that even careful reflection of the pros and cons yields no clear indication, it is the patient's fear or courage which decides the issue. In most cases, however, surgical prognoses have a much more solid foundation than those of medicine because, as we have seen, surgeons have resolutely rejected the extremely individualistic notion that every case is different.
Still, even medicine is beginning to catch up. Where statistics were formerly frowned upon as being unworthy of a true doctor, British attempts,^ in particular, have led to a new concept in medicine: the Medicine of Probability, ^ whose task it is to classify pathological symptoms so accurately that individual diagnoses and prognoses can be made almost by rote.
Though British doctors are still far from having attained this objective, the Medicine of Probability promises to clear medical prognosis of its present state of utter confusion. The 6000 experimental data that have been compiled for different diseases so far, have already proved their great prognostic worth.
^ A. Bradford Hill: Principles of Medical Statistics (London 1950). 2 A. Fidler: JVhither Medicine? Part 2: The Medicine of Probability (London 1946), pp. 41-87.