View report page

109
[1938
The number of houses having three or more cases is so small that the excess
of the observed over the expected number cannot be regarded as significant. However,
in the case of Bristol, as we have already noted, this excess can be fully
explained by the segregation in certain streets of a number of persons whose ages
range from 55 to 75. When we bear in mind that these tables relate to 129,000
inhabited houses it can scarcely be disputed that the incidence of the disease is in
accordance with the laws of chance. If there were any appreciable tendency for
the disease to spread from one person to another in the same house or to people in
adjacent houses a very considerable increase in the number of multiple cases would
be expected. The figures for Islington relate to a considerable area of the
Administrative County of London, which has similar environmental circumstances
to, and merges imperceptibly into, the remainder of London. It therefore seems not
unfair to assume that what holds good for this borough will probably hold good for
the remainder of London, and not improbably for other areas.
Reverting to the paper by Leyton and Leyton (1937) the authors dealt with a
mixed urban and rural area, the latter of which cannot be compared with a metropolitan
borough, where the environment is comparatively standardised. In the
rural area there are possibly variations in water supply, drainage, and other factors
in the various areas. But if we consider the figures given for the city of Oxford, 123
only in number, it seems fair to suggest that in the light of the results obtained at
Islington and elsewhere no definite conclusions can be drawn from such acomparatively
small series of cases ; moreover, it is a mistake to assume that a chance
distribution is an even distribution unless a very large number of cases are considered.
Summary.
1. The house-to-house distribution of 2,464 cases of cancer occurring in Islington
during the years 1932-6 has been analysed.
2. It is shown that this distribution corresponds very closely with the expected
distribution calculated according to the laws of chance, the criterion being the
number of houses in which multiple cases occur.
3. This correspondence is no less striking when the Islington figures are combined
with those obtained (by P. Stocks) at Bristol in the years 1922-7 and at
Worcester in the years 1921-30.
4. It is argued that theories based on the grouping of cases in particular areas
are often misleading.
5. It is suggested that the distribution of cancer in general follows the laws of
chance so closely that it is superfluous to introduce any other hypothesis to account
for it.
6. These facts are some evidence against an infective or contagious theory of
cancer spread, and invalidate any remaining belief in "cancer houses."
References.
Leeson, J. p. (1908). Practitioner, 80, 256.
Leiper, R. T. (1926). Rep. Brit. Emp. Cancer Camp, for 1925.
Leyton, G. B., and Leyton, H. G. (1937). British Medical Journal, 2, 378.
Sambon, L. (1926). J. Trop. Med., 29, 233.
Stocks, P. (1935). J. Hyg., 35, 46.
* The full paper was printed in the British Medical Journal Supplement, 3/9/38, and I am
indebted to the Editor for permission to extract the salient points.—G.C.T.