raised in the same proportion in order to bring it into comparison with the death-rate of England and Wales-i.e., it must be increased in the proportion of 20.66 to 21.27. But the 21.27 fraction =10296. This, then, is the factor for correc 20.66 tion for age and sex-distribution by which the recorded deathrate of Brighton must be multiplied in order that it may be comparable with that of England and Wales. It may be objected that the mean age and sex-distribution of the decennium 1871-80 is assumed in the above calculation to hold good for every year of the succeeding decennium. This, however, can be proved to be approximately accurate, inasmuch as in populations whose chief occupations remain the same the age and sex-distribution after a time assumes constant proportions. It may be objected again that the factor or correction figure, which is accurate when applied to the rate of an entire decade, may not be equally accurate for each year of that decade. But on selecting for comparison two years, one extremely hot and therefore dangerous to the young, and the other extremely cold and dangerous to the aged, the Registrar-General found that the error due to the use of the decennial correction figure as a constant, instead of a special correction figure for each year, was so small that it might practically be disregarded. By multiplying the recorded death-rates in column 3 (page 92) by the factors for correction, we obtain the corrected death-rates given in column 4. These are the death-rates which would have been recorded in each town had its population been identical, so far as age and sex-distribution are concerned, with the population of England and Wales. Trustworthiness of General Death-Rates. The following remarks of the Registrar-General (Annual Summary, 1883) on this point are apposite: "It may naturally be asked, of what use are the general death-rates, as usually given, if they cannot be accepted without further and considerable correction? In the first place, if the death-rate in any given town or other area in one year be compared with its death-rates in other years, no correction is required; for the age and sex-distribu tion in an individual town or other area remains practically constant; and, secondly, although it is doubtlessly true that the general death-rates of towns or other areas cannot safely be used for accurate comparison between such towns or areas in respect of healthiness without further correction, yet they serve as a very valuable approximate indication; for if the column 3 be compared with column 4, it will be seen that whether the towns be arranged according to their recorded or according to their corrected death-rates, the order will scarcely be changed. The correction simply alters the amount of difference between the towns, leaving the position in which they stand to each other but slightly changed." In the table on page 92 it will be seen that only in two of the twenty-eight towns, Plymouth and Norwich, is the corrected death-rate lower than the recorded death-rate. In all the other towns a correctional addition to the recorded death-rate is required, varying in amount from 0.50 to 3.28. If the death-rates in the counties are corrected for age and sex-distribution, so as to conform to the standard of England. and Wales, the corrected rate in the greater number of the counties is somewhat lower than the recorded rate. In the table given at page xviii. of Dr. Ogle's Supplement to the Registrar-General's Report for 1871-80, only ten counties have the corrected rates higher than the recorded rates. Of these ten, all, except the extra-metropolitan parts of Surrey and Middlesex, are industrial counties, where the death-rate even before correction is very high; while all the agricultural counties are comprised in the majority, and have their death-rates, which are already low, lowered by correction. Before correction the extreme limits of variation in the death-rates of different counties lie between 16.77 and 25.17, while after correction they range from 15.97 to 26-87 per 1,000. As a general rule, it may be stated that correction raises the rates of towns and industrial centres, but lowers the rates of rural districts. Another instance of the importance of correction for age and sex-distribution may be quoted from Dr. Ogle. In 1881 the general death-rate in England and Wales was 18.9 per 1,000 of all ages, while the general death-rate in France was 22·0, i.e. 3.1 higher than England. But had the age distribution of the French population been identical with that of the English population, the French general death-rate would have been 20.9 and not 22.0. Thus of the 31 difference between the two rates, 2.0 was due to difference of health condition, and 11 was due to difference of age distribution. = CHAPTER VII. MALE AND FEMALE MORTALITY AT DIFFERENT AGES. Death-Rates at Age-Groups.-False Methods of Estimating these.-Infantile Population.-Infantile Mortality.-Mortality in each Month of First Year. Causes of Infantile Mortality.-Insurance of Children.—Mortality of Illegitimate Infants.-Female Mortality.-The Standard Million. N the last chapter, the disturbing influence of varying age In chaptention of the population on the general mortality has been discussed, and the method by which the fallacy involved can be avoided has been described. We have also pointed out that the relative position of the twenty-eight great towns is scarcely changed, the correction simply altering the amount of difference between the towns, but leaving their relative position practically the same. Death-rate at Age-groups.-A statement of the death-rate at various groups of ages per 1,000 living at these ages, is quite trustworthy for comparison with other towns and districts. To obtain this, it is necessary to know 1. The population at different age-groups. 2. The deaths at different age-groups. The latter are obtained from the local registrar's deathreturns; the former may be ascertained in the case of each urban and rural sanitary authority by an application to each group of ages of the same method of estimation for increase of population as that described on page 6. For the population at all ages the same rate of increase at each group of ages, as in the previous intercensal period, is assumed to continue during the present decennium. Thus, if the numbers living at each age-group in 1871 and 1881 are known from the census returns, the calculation is short. If, however, in the 1871 census, the age-constitution of the population in question was not given, a subsidiary calculation is required. It is assumed that the age-distribution does not vary between the two census enumerations, an assumption practically correct when the prevailing industry in a given neighbourhood has not altered. Then having given the total population in 1881 and in 1871, and the population distributed according to age and sex in 1881, the finding of the distributed population of 1886 is a mere question of proportion. Thus if the total population of a town in 1881 is 107,546, and the number living aged 10-15 is 10,741, while the total population in 1871 is 92,481, to find the number living aged 10-15 in 1886. The calculation will be in three stages. (1) Find rate of increase per unit of population from 1871 to 1881. And log. (1+r)= (log. 107546-log. 92481). And r rate of increase per unit of popula tion='015. (2) Find from this the total mean population for 1886, i.e. after the lapse of 5 years. If P=mean population of 1886. P=107546 (1.015). Hence P=116363. (3) Find in this population the number living aged 10-15. If x this number, Then 107546: 116363 :: 10741 x. Therefore x=11625. |