tality from renal diseases, confirming the general medical opinion that lead-poisoning causes renal disease. In the four years 1879-82, only 229 deaths of males were ascribed to lead-poisoning, though this number would doubtless be greatly increased were it possible to include the deaths from renal nervous and other affections produced by lead-poisoning. The 229 deaths were distributed as follows: The annual death-rates from lead-poisoning among males over fifteen years of age, as deduced from a comparison of these figures with the census returns of 1881, are as follows. Reliance cannot, however, be placed on rates calculated from so few data: For further details on all the questions discussed in this chapter, which have a great practical importance, the reader is referred to Dr. Ogle's invaluable report, pages xxi.-lxiv., which ought to do much to stimulate preventive measures in the case of unhealthy occupations. CHAPTER XI. MORTALITY FROM ZYMOTIC DISEASES. Methods of Estimating.-Zymotic Death-rate.-Measles.-Scarlet Fever.Case Mortality.-Relation of Scarlet Fever to Age and Sex.-Geographical Distribution of Scarlet Fever and Diphtheria.-Diphtheria.-Whooping Cough.-Fever.-Enteric and Typhus Fever.-Diarrhoea. WE E have in the preceding chapters shown the effect produced by various social and other conditions on the general mortality. In the next place we must consider the mortality from individual diseases and from groups of diseases. In the first place, as to the methods of stating this mortality. This will vary according to the object in view. (1) The first plan, useful only for purely medical purposes, consists in stating the proportion of deaths to persons attacked with any given disease. This method is of importance to medical men, as from the results thus obtained deductions are drawn concerning the effect of a particular line of treatment. We may point out the liability of this method to error before passing on to the methods of statement usually adopted by vital statisticians. (a) The number of facts manipulated is often so small as not to warrant exact conclusions. I have recently investigated an outbreak of enteric fever due to infected milk, in which 21 cases occurred. Of these 15 were treated either at home or in a public hospital, while 6 were treated in another public hospital. Of the first 15 cases none died, while 2 of the last 6 were fatal. The difference in fatality was almost certainly due to accidental causes, and any inferences as to skill of treatment would be altogether fallacious. Speaking generally, therapeutical results or etiological theories, advanced on the strength of percentages from a small number of cases, must be accepted with caution. (b) The two groups compared may have a different age and sex constitution, and it is a notorious fact that the death-rate from nearly all diseases is largely influenced by these factors, and especially by age. This is seen in the following table from Murchison on Continued Fevers. Mortality per cent. of Males and Females Attacked at Different Ages. Age. 0-5. 5-10. 10-15. 15-20. 20-25. 25-30. 30-35. 35-40. 40-45. 45-50. Typhus (Males 12.50 11-17 11-12 12-76 23-30 17:17 32-69 26-04 28.12 33.33 Fever Females 6.69 3.59 2.28 4:46 10:33 15-17 20:55 25-92 30-79 42-54 Of course, by splitting up the deaths at all ages into groups at different ages, the number of individual facts at each age is reduced, and thus the fallacy due to paucity of data is introduced. This, however, is preferable to the lumping together of two sets of facts which are not comparable with each other. In stating percentages the number of facts on which they are founded should always be given, as their trustworthiness can then be gauged. (c) In the case of infectious diseases, the character of the particular epidemic must be taken into account. One epidemic of diphtheria may differ in virulence from another, and comparison between methods of treatment in the two epidemics would be consequently largely vitiated. Similarly, it would be unfair to compare the treatment at the beginning of an outbreak of cholera, when 50 per cent. of those attacked usually die, with the treatment in the later and milder period of the epidemic. (2) The mortality from any given disease or group of diseases may be stated as a proportion to the deaths from all causes. This method is, however, essentially fallacious, as it constitutes a ratio between two factors, of which both are variable, viz., the mortality from the specified disease, and the mortality from all causes. Dr. Ransome gives the following example of its fallacious character. Suppose a town of 100,000 with 2,000 annual deaths, of which 500 are caused by phthisis. Here the general death-rate is 20 per 1,000; the death-rate from phthisis is 5 per 1,000 living, and the proportion of deaths from phthisis to the total deaths is 250 to 1,000. In another town having the same population the total deaths are 4,000, and therefore the death-rate 40 per 1,000 inhabitants; the deaths from phthisis are 1,000, and therefore the death-rate from phthisis is 10 per 1,000; but the proportion of the phthisical to the total mortality is 250 to 1,000, as before. In the second town, therefore, there is by the latter test apparently no worse condition, so far as phthisis is concerned, than in the first, though matters are really twice as bad. The following table, abstracted from the local report of a medical officer of health, gives another instance of this fallacious method:: The zymotic and general death-rates are in this table. accurate indications of mortality; but if the two first columns are compared, it will be seen that the proportion between the two gives no indication as to their relative mortality. Such relative mortality might be increased either by a diminution in total deaths or an increase in zymotic deaths, though the inference to be drawn in the two cases would be very different. For large communities, or for the whole country, when fallacies due to varying age and sex constitution can be eliminated, it is often convenient to know the percentage of deaths due to different causes. The following table shows the percentage of deaths from different groups of diseases in England and Wales during 1886; and in a parallel column the rate of mortality per million of the population. |