View report page

23
There is nothing in these figures to support the hypothesis that the doctor is a
carrier of puerperal infection, since in the boroughs where obviously the greater
proportion of births have been delivered by doctors or medical students the
puerperal fever mortality is lower than in the group where midwives' cases are in a
majority, without any compensating excess in deaths from other causes.
The Registrar-General in his annual report for 1923 (Text p. 96 et seq.) directs
attention to the remarkable increase in maternal mortality from sepsis in the latter
months of 1919 and the beginning of 1920, and shows the rates in relation to the
demobilisation records of 1919. He observes:" It has been suggested that the
explanation of this correspondence between demobilisation and puerperal sepsis
resulting from conceptions of the same date may be provided by the liability of the
site of gonorrhœal infection to secondary streptococcal invasion." The quarterly
rates which he gives on page 98 show, however, that the increase in mortality had
already commenced in the second quarter of 1919 when demobilisation could not
have entered into the question, and, further, the increase in mortality shown to
have occurred in 1914 cannot be associated with any event of a similar kind. The
epidemiologist will in fact hardly accept any hypothesis of the kind suggested as
accounting for the facts in view of the observation of Dr. G. B. Longstaff in a
paper read before the Epidemiological Society nearly 50 years ago, in which he
pointed out the close inter-dependence of movements in the mortality from
erysipelas, puerperal fever and scarlet fever ("Studies in Statistics," Dr. G. B.
Longstaff, p. 310). For the purpose of illustrating the relative variations in
puerperal fever mortality and erysipelas deaths in England and Wales a diagram is
shown facing page 25. It will be seen from the upper contour that during the
years 1911-14 and 1921-27 the fluctuations in mortality of the two diseases show
remarkable similarity and the variations in puerperal sepsis mortality of the intervening
years must therefore be regarded as due to the causes, whatever they may be,
which have determined the variations of the rates in preceding and succeeding years.
In the lower contour the deaths from erysipelas among women aged 15-45,
above and below the mean, for the years 1913-27, are shown in relation to the deviation
of the actual puerperal fever mortality from the rate which would be expected if the
puerperal fever deaths stood in a constant relation to the deaths from all other
causes in childbirth, in which of course no zymotic influence comes into question:
thus the mortality per 1,000 births in childbirth from causes other than puerperal
fever in England and Wales during 1913-1917 averaged 2'577 while the puerperal
fever rate was 1.461, or .567 that of the former: and if the mortality from other
causes in each year be multiplied by .567 rates approximating to the actual puerperal
fever rate will be obtained. The divergences of the actual rates from these
calculated rates are shown by the contour in the lower section of the diagram to be
similar to the divergence of the erysipelas deaths in each year from the average
deaths for the same period: a result consistent with the view that the two diseases
have a zymotic association, in which other causes of maternal mortality have no part.
Dr. Longstaff showed (loc. cit.) that the annual rainfall influenced in a general
wav the prevalence of the zymotic diseases with which he dealt in his paper.
Fluctuations
in annual
sepsis rates.
It will be observed that the maternal mortality rates have been calculated
per 1,000 live births still births and miscarriages have been ignored. With regard
to still births, it is not found that social condition influences their proportion to
live births in any considerable degree so far as can be observed in the metropolitan
boroughs, nor has their proportion to live births varied since 1910 sufficiently to
affect in any measure the comparability of the rates shown. With regard to
miscarriages no information is available. The percentage of pregnancies terminated
by stillbirth or abortion is variously stated as being from 14 to 30 per cent. (Problems
of Population and Parenthood, p. lxxx). The figures for Aberdeen show that in
1918-1927 of 252 deaths associated with childbearing 109 or 43 per cent. occurred
after stillbirth or abortion or before parturition.
Concluding
observations.