a particularly simple matter. The patient must either undergo the
operation or not, and the proportion of cases in which the necessity is
doubtful is very small. Further, he either recovers or dies--for the
operation could hardly be in itself unsuccessful, nor the surgeon in
doubt as to whether the patient had recovered or not. Many operations,
however, are not of this order. They may be performed for conditions as
to the exact nature of which the surgeon is uncertain, and for symptoms
which may be only partially relieved. Thus, the removal of the appendix
for Appendicitis may be most urgently necessary for the saving of life
in one case and may be a matter of convenience for the relief of more
or less indefinite symptoms in another. Further, what one surgeon calls
appendicitis another may not. One surgeon may have every appendix
that he removes submitted to skilled pathological examination before
he accepts the case as one of appendicitis and places it among his
statistics. Another may be quite content with naked-eye appearances of
the nature of which he alone is witness, judge, and reporter. It is,
therefore, clear that any collective statement as to the results of
such an operation must be cautiously scrutinized before conclusions of
the slightest scientific value can be drawn from them.
[Illustration:
FIG. 135. DEATH-RATE FROM CANCER OF THE TONGUE. It will be observed
that it is not a common cause of death till about 45 years of age,
but that it then increases rapidly to fall again in both sexes in old
age. These features are clearly related to various factors in the
causation of the condition. One of these is certainly Syphilis, which
is most frequently contracted between 20 and 30 and more often by men
than women. The so-called ‘tertiary’ effects of this condition, some
of which lead to Cancer of the Tongue, do not usually make themselves
felt, however, for many years after infection. Contrast Fig. 136 and
Fig. 137. ]
There is a common and rather foolish saying that ‘Statistics may be
made to prove anything’. This is true, but it is true only in the sense
that _evidence_ may be made to prove anything. The matter turns on the
questions, firstly whether the evidence is of a good or a bad order,
and secondly whether the investigator is in a good or bad position to
interpret the evidence. A statistical statement may be well or ill
founded and well or ill interpreted, but statistical statement is, in
fact, the only scientific method open to us for presenting long series
of data. The conclusions to be drawn from those data, though sometimes
evident and easily elicited, at other times demand specially skilled
and specially trained interpreters. Moreover, to be of value to others,
such interpreters must also be skilled in expression, so that the main
body of those who have no statistical training may be in a position to
understand the essential elements in their conclusions. In no medical
department is literary power of greater importance than in that which
deals with statistics. Thus has arisen the small but highly important
class of medical statisticians. The rise of medical statistics into a
vocation places the crown on Medicine _as a science_. It is not given
to many medical men to be proficient in this department. But the duty
lies on all medical men, and indeed on all citizens, to appreciate
the value of this study and to seek to appraise its simpler and more
established conclusions.
It is remarkable how frequently a straightforward statistical statement
may remove a false impression, even when the impression is based on
evidence not of a wholly unscientific character.
[Illustration:
FIG. 136. DEATH-RATE FROM CANCER OF THE LIP. It will be observed that
this curve resembles in form that of the death-rate from Cerebral
Haemorrhage as shown in Fig. 137, but differs from that of the
death-rate from Cancer of the Tongue as shown in Fig. 135. The chances
of dying from Cancer of the Lip are negligible till middle age is past
and then increase progressively throughout life. In the causation of
Cancer of the Lip Syphilis is not an important factor. On the other
hand the continuous irritation of pipe-smoking, which acts not at one
age but throughout life, has to be considered as a causative element.
Hence the resemblance to Fig. 137 rather than to Fig. 135. ]
[Illustration:
FIG. 137. CHART OF DEATH-RATE FROM CEREBRAL HAEMORRHAGE AND ALLIED
STATES. These conditions are extremely rare in the young, but among
the commonest causes of death in later life. The liability to them
increases progressively to extreme old age. This is explained by the
fact that Cerebral Haemorrhage, etc. follows on the rupture of a
blood-vessel in the brain and the rupture of the vessel is conditioned
by the hardness and brittleness of its coat. The hardness of the
arteries increases progressively in later life, whence the saying ‘a
man is as old as his arteries’.
]
For example the increase in the incidence of deaths from Cancer has
often been emphasized. But Cancer is a disease of advancing life. The
age distribution of the death-rate from many forms of Cancer is closely
parallel to that of certain other forms of senile disease (Figs.
136 and 137). Now the age constitution of the population of most civilized
countries is altering in the sense that the proportion of the elderly
and aged is constantly increasing (Fig. 134), so that some increase in
the Cancer incidence must be expected. Moreover the appearance of some
increase in the incidence of Cancer is due to improved diagnosis. How
far there is a real increase, when these factors have been taken into
account, is still somewhat doubtful. It must always be borne in mind
that a relative decrease in the proportion of deaths from _any_ cause
must automatically increase the proportion of deaths from other causes.
Again, there is no doubt of the fall in the death-rate in England
and Wales from ‘Phthisis,’ or pulmonary tuberculosis, during the last
fifty or sixty years. There is also no doubt of the effect both of
bad housing and of urban conditions in inducing a susceptibility to
chest disease in general and to pulmonary tuberculosis in particular.
Further, there is no doubt that the rural population suffers less
from pulmonary tuberculosis than the town population. These matters
of common medical knowledge have naturally led to the conclusion that
the rise of the great towns has led to a great increase of pulmonary
tuberculosis, and that this increase has been remedied by the improved
housing and sanitary conditions of the last generation. A study of the
statistical evidence, however, negatives this view. The rise in the
proportion of deaths from pulmonary tuberculosis took place before the
Industrial Revolution. Moreover, the proportion began to fall long
before the campaign against tuberculosis could affect the issue. The
history of pulmonary tuberculosis may, in fact, be regarded as that of
an ‘epidemic’ outbreak, extending over about 100 years, of a disease
which has always been endemic and remains so now that the epidemic is
past.
[Illustration:
FIG. 138. CURVE SHOWING PERCENTAGE OF DEATHS FROM PHTHISIS to total
deaths from all causes in London over a period of 200 years. It will be
seen that the percentage begins to rise definitely about 1730 and to
fall definitely about 1830. This state of affairs may be pictured as an
epidemic lasting about 100 years. ]
These points are well brought out in the accompanying diagram (Fig.
138). The fall in the proportion of deaths from Phthisis expressed
there gives rise to further considerations. It might seem that the
statement that the proportion of those who died from phthisis was
diminishing left in itself no doubt that the disease was less prevalent
than formerly. This, however, is not the case. Phthisis is more liable
to affect those under forty-five years of age than those who are
older. Now the proportion of the population that is under forty-five
is steadily diminishing (Fig. 134). This is one of the results of the
steadily diminishing general death-rate (Fig. 96, p. 196). Therefore
the proportion of the more susceptible to the less susceptible is
diminishing. It might have been the case (though it is not) that the
ratio (more susceptibles)/(less susceptibles) was not only decreasing
but was actually decreasing more rapidly than the ratio (total
deaths)/(deaths from phthisis). Had this been so, the conclusion would
have been justifiable that the fall in the proportion of deaths from
the disease did not correspond to any decrease in its infectivity. In
fact, however, the prolonged high mortality from phthisis and its later
rate of fall do suggest the former prevalence of a more virulent type
of the disease over a long period, in other words something of the
nature of a prolonged epidemic.
This conclusion leads us to the conception of the nature of an
epidemic. To gain some conception of the ideas involved in that word,
we must glance back in history.
From the time of Hippocrates onward the subject of Epidemic outbursts
of disease has drawn the attention of physicians. A writer in the
_Hippocratic Collection_ thought he could perceive an association
of symptom-complexes with each other and with the weather. In the
great work _Epidemics_, to which the name of the Father of Medicine
is attached, such a view, known as that of ‘Epidemic Constitutions,’
is set forth. The view was revived by Sydenham in the seventeenth
century and has given rise to a vast literature extending to our own
time. In the eighteenth and nineteenth centuries the attempts of the
investigators of vital statistics to place the leading events of life
in a form capable of exact analysis (pp. 166-68) focused attention
on the search for a mathematical expression for the rise and fall of
epidemic diseases.
The first successful attempt to describe epidemics along these lines
was made by William Farr (1807-1883), an official in the office of
the Registrar-General in London, and one of the greatest of all
epidemiological thinkers. His first publication on the subject was
in 1840, and had reference to the recent outbreak of small-pox, in
which more than 30,000 had died in England and Wales. It was his merit
to observe that the successive decreases in the number of cases in
successive equal periods during the decline of the epidemic correspond
to the successive increases in the number of cases during successive
equal periods of the rise of the epidemic. In other words, he observed
that the rise and decline of an epidemic tend to be mathematically
symmetrical.
Farr’s suggestion that epidemics are liable to follow the lines of
regular mathematical rules drew little attention at the time, but in a
later year it led to a most remarkable and striking prophecy. At the
end of 1865 Cattle-plague broke out in England. Week by week the number
of cases increased. In the fourth week of February 1866 the responsible
Minister, in a speech in Parliament, gave a very gloomy account of
the state of affairs, expressing the belief that the devastation
would be far beyond what had yet been encountered. Farr, however, had
been watching the returns, and had been applying his rule to them. He
thereupon made a public pronouncement of his belief that at an early
date the outbreak would reach its maximum and would then decline. The
outbreak did, in fact, very closely follow the course which he had
predicted by reasoned calculation. Farr even prophesied the number of
cases that would occur week by week. His prophecy was near the truth.
During the years that followed Farr’s prediction his views were applied
with success to a variety of epidemic conditions. The regular form of
the development of the epidemic was found to apply in certain outbreaks
of typhus, measles, and other conditions.
Farr’s law was more exactly expressed by him in 1868. It remained,
however, simply a mathematical law, a rule of which the underlying
cause was not apparent. It was soon observed that his law applied to
many but by no means to all epidemics. Moreover, it was perceived that
the actual figures which he gave for his epidemic of 1840 resembled
those of certain other epidemics in that they could be fitted with
greater or less exactness to a well-known mathematically described
curve, known as the ‘normal curve of error’. We need not discuss the
mathematical foundation of this curve, which is shown in two variants
in Fig. 139. For our immediate purpose it is enough to observe that it
rises gradually at first, but then more steeply, that the steepness
decreases after a while, and then the curve begins to decline again, as
it rose. We note that it is symmetrical.
[Illustration:
FIG. 139. THE NORMAL CURVE OF ERROR, shown in two types made with the
same formula but with different constants. This curve has been shown
to be similar to that representing the incidence of cases in some
Epidemics.
Vertical lines are drawn from two pairs of symmetrical points. The
continuous lines refer to the higher curve, the broken lines (from the
points of intersection of the two curves) refer to the lower curve. The
lines will be seen to divide the curves into three parts. This division
is of such a character that the sum of the two lateral areas is equal
to the central area for each case. ]
[Illustration:
FIG. 140. Curve of monthly number of deaths from Small-pox during an
epidemic at Warrington, Lancashire, in 1743.
FIG. 141. Curve of weekly number of cases of Scarlet Fever registered
during an epidemic at Glasgow in 1892.
Both curves are fitted to the theoretical epidemic curve, and are
modified from Brownlee. The curves are in both cases explained on the
assumption that the infectivity, having reached a high point at the
beginning of the outbreak, decreases thenceforward in geometrical
progression. ]
When we are dealing with living beings or are dealing with things that
may indefinitely approximate to a mathematical rule, but never entirely
fit it. Especially when the living beings are also human beings, with
their infinitely complex relationships, various factors are present
which interfere with the exact application of mathematical findings.
Nevertheless, the theoretical form of the epidemic is an extremely
useful framework into which actual epidemics may often be fitted, with
greater or less exactness. In the accompanying figures (Figs. 140-141)
are adduced cases of greater exactness. There are, however, many cases
in which an ‘outbreak’ does not seem to fit the simple theoretical
curve at all. Examination of such curves has in some cases suggested
that we have not one epidemic or disease but two or more to deal with.
In some cases it has been possible to analyze the outbreak on the basis
of two or more theoretical curves, suggesting in fact two or more
outbreaks of similar but not identical causation (Fig. 142).
[Illustration:
FIG. 142. THE CURVES OF SOME EPIDEMICS, which do not follow the
theoretical curve, may be analyzed as compounded of two or more
epidemics, each of which accords individually to the mathematical rule.
Thus ‘Summer Diarrhoea’ is a seasonal disease very fatal to infants
in England during the hot months, July and August. The angular curve
shows the average daily incidence of deaths from this disease in London
during the fifty-three years 1850-1903. It can be analyzed into two of
the theoretical epidemic curves.
Each reading of the curve, calculated from the actual cases of ‘Summer
Diarrhoea of Infants’, can be divided into two, as indicated in the
step-like readings, one dotted and the other continuous. These accord
beautifully with two theoretical curves, thus indicating not one but
two recurrent epidemics. It thus seems probable that two separate
sources of infection are confused as ‘Summer Diarrhoea of Infants’.
]
What can be the causative element which constrains the incidence of a
disease in a population to follow mathematical rules? An answer was
provided by John Brownlee (1868-1927), the late statistician to the
Medical Research Council of England. The leading fact about an Epidemic
is that it rises to a maximum, falls, and then dies out, and that the
curve representing the number of new cases in a series of equal and
consecutive periods of time throughout the Epidemic is symmetrical. In
practice the decline is usually a little slower than the rise. This
is sometimes, at least, due to better observation and record of the
later cases. Now why does an Epidemic die out? The possible reasons
may be reduced to three. Firstly, the end of an Epidemic may be due to
the exhaustion of susceptible persons in the population. That is to
say, all the survivors are immune, either being so by nature or having
become so by having contracted the disease and recovered. Secondly, it
is conceivable that the liability to the disease should be decreased,
not by rise in the proportion of immunes, but by externally acting
causes, as, for instance, by rise of seasonal temperature, which would
provide conditions under which the organism loses its infectivity.
Expressed in older language, this is to say that the ‘Epidemic
Constitution’ (p. 342) has changed. Thirdly, the infecting organisms
may, of their own inner nature, lose their infectivity. The second
factor may act in special cases, but may be disregarded except in those
cases. We are, therefore, left with the first and third.
Now it is possible to construct curves that would correspond to the
exhaustion of the supply of susceptible persons by continuous increase
of the proportion of those who become immune either by taking the
disease or by dying. These curves, however, have the character that
their descent is more rapid (and neither as rapid nor less rapid) than
their ascent. It is the merit of Brownlee to have suggested that the
actual curve of the Epidemic corresponds to a known though very little
understood biological phenomenon, namely change in the infectivity of
the invading organisms. The simplest expression of his discovery is
that the loss of infectivity of these organisms is approximately in the
ratio given by a geometrical progression. That is, if the infectivity
of the Epidemic be _m_, and at the end of a unit of time _mg_ (when _g_
is less than unity), at the end of a second unit of time it will be
_mg_^2, and at the end of the third _mg_^3, and so on. Assuming this
to be a fact, the course of Epidemics would follow the curve of normal
error (Fig. 139).
Of late years it has been possible to institute artificial epidemics in
a series of animals under control conditions. Such experiments must,
in the nature of the case, cover a large number of years, but they bid
fair to throw much light on the nature and progress of human epidemics.
These results seem to show, what was believed on other grounds, that
in the case of highly infective disease, to which, in any population,
there are many highly susceptible, isolation of declared cases has
little or no effect on the course of the Epidemic. Such diseases are
Scarlet Fever, Measles, Influenza, &c. Moreover, the experimental
epidemics seem to confirm the conclusions of Brownlee that in some
cases at least the course of the epidemic is determined by biological
changes within the parasitic beings that cause it.
Thus in the end our health, our lives, and indeed the continuance
of our civilization may well depend upon a factor which is outside
ourselves. For reasons which we know not, the pullulating billions
of living things which are around us, upon us, within us, take up a
virulence which before they had not, and after a time they lose that
virulence to become as they were before. The world is devastated by
an outbreak of Plague, of Cholera, of Influenza. But how and why
the organisms that carry these diseases should acquire a new and
more deadly infectivity lies among the secrets yet locked within the
living cell. Life--the life of the Cell, of the Bacterium, of Man
himself--remains among the _Arcana Naturae_. These are the secret
things that in their essence--which is Life--remain and will remain
behind the veil. From such cells we came, through such cells we shall
return. As to what is the force which starts these processes on their
way, we are as ignorant as children, and must remain so, in essence,
till we understand the nature of the processes of coming into being and
passing away. So Medicine must end where she began, quaking before the
Mystery of Life, a Mystery which could only be resolved if we could
express Mind in terms simpler than itself. If this could be done the
veil that is cast over all flesh would indeed be rent. But the author
of this work believes that the hope of this is vain and that we are
here in the presence of one of the ultimate things.
EPILOGUE
We have now traced various movements in Medicine throughout the ages
and have seen how all the sciences in turn have been made to bring
their tribute to the alleviation of suffering. We have seen especially
how the consideration of disease as a whole, and of the health of
peoples as a whole, has introduced a new view in the handling of
disease. Health is a public asset, and its promotion has now been
recognized as a public duty. There are undeniable disadvantages in
placing officers of the State in control of the personal liberties of
its citizens, but, on the whole, the advantages, in matters of health,
have outweighed the disadvantages. Only a professed pessimist or a
crotchety reactionary could deny the gains to humanity from the passage
of preventive measures from private into public hands.
There is another side of the picture which we have need also to
consider. The advances in Medicine and the advantages that have accrued
therefrom have been entirely the result of the application of the
rational method of observation and experiment. To control Nature we
must above all things understand Nature. Neither the conception of
Nature as the kind old nurse nor the conception of Nature ravening red
in tooth and claw will stand. Least of all can we tolerate the picture
of Nature as a bountiful mother. If we go to her asking something for
nothing, she (far from bountiful) will give us little but what we have
given her, and to him who but begs she gives no more than a beggar’s
portion. It is thus that she has served the magician and the wizard,
who think they can compel her to give them all things by their paltry
charms!
The amount of human labor and ingenuity that is now being thrown
into the investigation of Nature is almost incredible even to men
of science. Some conception of the enormous and unreadable bulk of
scientific literature may be gained by a glance at the _International
Catalogue of Scientific Literature_. This gives the _titles alone_
of original articles in the various departments of physical science.
These titles for the year 1914 alone occupied seventeen closely printed
volumes! The rate of publication has accelerated considerably since
then. There are now about 25,000 periodicals devoted to scientific
publications! There are very few departments of science which do not
have some bearing on Medicine. It is evident that no human mind can
possibly compass even a year’s output of this material.
And yet it is not the bulk of writing on Science that forms the only
or even the chief deterrent to the general comprehension of its
principles. The mass of scientific detail has always been beyond the
power of one mind to grasp. But as we have traced Rational Medicine
through its long course in Antiquity and the Middle Ages to its
debouchment on to our own time, we have found not only a more difficult
but also a new situation. In approaching our own age we have found ever
more difficulty in discussing Rational Medicine as a single channel of
thought. It spreads into a Delta, of which, though the many mouths may
inosculate, yet the tendency seems to be for an ever wider divergence.
This diffusion, induced by increased specialization, cannot go on
for ever without defeating the very objects for which specialism was
invented.