Journal of Hygiene

Research Article

Diphtheria: A suggested explanation of the relative change in age incidence

E. A. Cheesemana1, W. J. Martina1 and W. T. Russella1

a1 From the Division of Epidemiology and Vital Statistics, London School of Hygiene and Tropical Medicine

The purpose of the investigation was to examine the reason for the relative age shift in the incidence of diphtheria. The disease has been represented as now concentrating on children of school ages. We analysed, for various quinquennial periods beginning with 1901–5, the statistics of the notifications and deaths according to age and sex in the London boroughs which we classified into four social classes where I to IV represented descending order in the social scale. The conclusions arrived at were:

I. The age shift as represented by the ratio of the mortality at ages 0–4 to ages 5–9 years which has taken place in London has been found to occur mainly in the poorer districts. The ratio in the residential districts has not altered appreciably during the thirty years under review (Table V).

II. The age shift is also indicated by the morbidity statistics for the three quinquennial periods, 1901–5, 1909–13 and 1919–23, as there was a significant increase in the mean age of attack of children under 15 years of age.

The case rate at ages 5–14 during 1929–33 as compared with that in 1901–5 has increased most amongst children in the lowest social group.

III. Each of the variables (a) the decrease in the birth-rate between 1911 and 1931 in each London borough—an index of the change in the size of family—and (b) the decrease in infant mortality—a measurement of environmental improvement—is positively and significantly correlated with (c) the corresponding change in the size of the ratio of the case rates at ages 0–4 to 5–9 years. The coefficients were:

S0022172400011803_eqnU1

but the partial correlation coefficients

S0022172400011803_eqnU2

indicate that the change in the size of family is the more important factor as the value of rac.b is fairly large and significant, whereas the correlation between the change in the ratio and the environmental improvement when the effects of the decreasing birth-rate are made constant—rbc.a—becomes statistically unimportant.

(Received January 11 1938)

Footnotes

Of the Medical Research Council's Statistical Staff

Metrics