COLUMBIA LIBRARIES OFFSITE HEALTH SCIENCES STANDARD HX64153819 RC261 .H6718 Some essential stati RECAP SOME ESSENTIAL STATISTICS OF CA1ICER MORTALITY - BY HOFFMAN RC.26/ Hb7/6 Columbia (Hnitier^ttp College of ^fjpsicianfi ano ££>urgeon£ ILibv&vp Digitized by the Internet Archive in 2010 with funding from Open Knowledge Commons http://www.archive.org/details/someessentialstaOOhoff Some Essential Statistics of Cancer Mortality Throughout the World By FREDERICK L. HOFFMAN, LL.D., F.S.S., F.A.S.A. NEWARK, N. J. PUBLICATIONS OF THE AMERICAN SOCIETY FOR THE CONTROL OF CANCER 105 EAST 22ND STREET, NEW YORK Bulletin 8 December, 1915 SOME ESSENTIAL STATISTICS OF CANCER MORTALITY THROUGH- OUT THE WORLD By FREDERICK L. HOFFMAN Chairman, Committee on Statistics, American Society for the Control of Cancer This reprint is a part of the Commemoration Volume, issued by the American Medical Association at its meeting in San Francisco, June 22 to 26, 1915, as c tribute to the medical sciences, which made possible the building of the Panamc Canal and the Panama Pacific Exposition. TZCUl Copyright, 1915 BY THE American Medical Association SOME ESSENTIAL STATISTICS OF CANCER MOR- TALITY THROUGHOUT THE WORLD In the furtherance of international cancer studies, it is obvious that the first prerequisite is trustworthy and comparable cancer mor- tality statistics. In course of time cancer morbidity data should become available through the cooperation of large hospitals and special insti- tutions for the treatment of cancer patients. An additional source of information should be the consolidated experience data of American and foreign life insurance companies. The amount of existing statis- tical information on the subject of cancer is truly enormous. In a forthcoming work on The Mortality from Cancer Throughout the World, I have made an earnest effort to bring together in a comparable form the information most likely to be useful in the furtherance of cancer research, so far as the statistical data can be of value in this direction. Aside from the scientific necessities of the existing situation, the statistical data are useful in advancing the nation-wide effort at cancer control through the required education of the medical profession and the laity as regards the menace of cancer and the urgency of improved methods of early diagnosis and of more qualified medical or surgical treatment of cancer in the early stages of the disease. It would serve no purpose to restate on this occasion the rather extended argument advanced in behalf of the utility of the statistical method in the study of the cancer problem in the work referred to, and it has therefore seemed better to limit the present discussion to a brief restatement of the salient facts of the worldwide study of the cancer problem on the basis of at least reasonably trustworthy and, in practically every case, official mortality returns. It may not be out of place, however, briefly to refer to the often repeated argument, that on account of inherent difficulties of exact diagnoses, the returns, particularly for earlier periods, are intrinsically untrustworthy and therefore not strictly comparable with the statistics of the present day. Nevertheless it has properly been pointed out by the late Mitchell Banks that, "while the diagnosis of cancer is prob- ably made much more frequently now than in former times, it required little skill to make the diagnosis at the time of the death of the patient. The diagnosis at such a time was by no means beyond the ability of even the rural practitioner of fifty years ago." The argu- 4 CANCER STATISTICS— HOFFMAN ment derived exclusively from necropsy records, that the clinical diag- nosis of cancer is often at variance with the facts, is in all probability decidedly overdrawn. Discrepancies between the clinical and anatomic diagnoses are much more common, in any event, in the case of many other important diseases than in that of cancer; and the evidence has not been forthcoming that in any considerable number of cases cancer deaths have been erroneously diagnosed as due to other causes, more or less allied with malignant disease. There exists a confusion of thought, caused in part by the misconception that the terminal diagnosis of can- cer which underlies the certificate of death, is assumed to be identical with the initial diagnosis of the disease, which, it hardly requires to be said, ever has been and ever is likely to be a most difficult problem con- fronting the physician at the bedside. On the basis of what is probably the most comprehensive statistical study ever made of a single disease in the history of medicine, I feel absolutely confident that the returns regarding the mortality from cancer are, in the main, trustworthy and strictly comparable for a considerable period of time. Subject to the foregoing observations, the accompanying tables are presented, with the further explanation that the -required amplification of the data and the discussion of the problems resulting therefrom are made conveniently accessible in the work previously referred to, which will be printed for gratuitous distribution in behalf of the cancer cause, by the Prudential Insurance Company of America. Table 1 indicates the approximate cancer death rate of the different continents and for the world at large, for the five years ending with 1912: TABLE 1.— MORTALITY FROM CANCER, REGISTRATION COUNTRIES OF THE WORLD, PERIOD 1908-1912 Total Continent Population Africa 9,041.866 America 382,293,573 Asia 272,814,962 Australia 27,939,842 Europe 1,425,516,942 Deaths Rate per 100,000 rom Cancer Population 3,018 33.4 251,535 65.8 148,447 54.4 20,276 72.6 1,082,704 76.0 Total 2,117,607,185 1,505,980 71.1 Table 1 is based on a total population of civilized countries, about 440,000,000, or about 25 per cent, of the entire population of the world, estimated for the year 1911. The statistics for the principal European countries are given in Table 2, for the period 1896-1910, to facilitate the convenient com- parison of the rate for three quinquennial periods. CANCER STATISTICS— EUROPEAN 5 TABLE 2.— MORTALITY FROM CANCER IN EUROPEAN COUNTRIES, PERIOD 1896-1910 , Rates per 100,000 Population — ^ Country 1896-1900 1901-1905 1906-1910 England and Wales 80.1 86.7 94.0 Scotland 77.1 84.8 99.7 Ireland 58.1 68.5 78.8 Norway 85.7 94.9 96.6 Denmark* 118.9 129.1 137.3 German Empire 70.8 77.7 84.2 Holland 91.9 97.8 103.5 Switzerland 127.4 128.3 125.9 Austria 68.9 74.7 78.3 Hungary 30.7 39.1 43.6 Italy 50.9 55.2 63.6 France* 97.3 92.1 102.7 All countries 69.1 74.2 81.0 * Cities only. According to this table, the European cancer death rate has increased from 69.1 during the first five years to 74.2 during the second, and finally to 81.0 during the third. The geographic distribution of cancer according to latitude for 130 of the world's largest cities, also for the period 1908-1912, is shown in Table 3. TABLE 3.— MORTALITY FROM CANCER IN CITIES ACCORDING TO LATITUDE, Jf JlKlUiJ lyus-iyiz Rate per No. of Degrees of Population Aggregate Deaths from 100,000 Cities Latitude 1912 Population Cancer Population 35 50 N.-70 N. 23,980,086 112,912,675 119,374 105.7 48 40 N.-50 N. 27,519,705 131,256.257 121,216 92.4 24 30 N.-40 N. 10,195,197 47,944,253 37,451 78.1 7 10 N.-30 N. 2,780,447 13,476,168 5,696 42.3 4 10 S.-10 N. 559,630 2,583,495 1,056 40.9 7 10 S.-30 S. 1,806,951 8,066,144 3,040 37.7 5 30 S.-40 S. 2,678,287 69,520,303 12,297,218 328,536,210 11,048 89.8 130 298,881 91.0 It is shown by Table 3 that the average cancer death rate for 130 cities was 91.0 per 100,000 of population; that the rate was highest in the most northerly inhabited latitudes, or that section of the globe which is comprehended between 50 and 70 degrees north latitude, and that for this section the rate was 105.7 diminishing to 92.4 for the cities located between 40 and 50 degrees, to 78.1 for cities between 30 and 40 degrees, to 42.3 for cities between 10 and 30 degrees, to 40.9 for cities between 10 degrees north latitude and 10 degrees south lati- tude, and finally, to 37.7 for cities between 10 and 30 degrees south latitude. In the most southerly inhabited belt, between 30 and 40 degrees south latitude, the cancer death rate again rose to 89.8, which is practically equivalent to the rate for 30 to 50 degrees north latitude. Table 3, therefore, would seem to warrant the important conclusion that cancer frequency is to a limited extent determined by latitude, which, of course, more or less represents the factor of climate and 6 CANCER STATISTICS— HOFFMAN weather .conditions ; in other words, cancer is excessively common in the temperate zone, moderately common in the medium zone and rela- tively rare in the torrid or semitorrid zone, which for the present pur- pose may be construed to include the belt between latitude 30 north and latitude 30 south. The possible relation of cancer mortality to the size of cities is of some importance in view of the unquestionable effect of the admission of non-residents to hospital treatment and the non-correction of the death rates by the redistribution of deaths according to the residence of the deceased. In the case of small cities with exceptional hospital facilities this factor, no doubt, is of considerable weight, but in the case of large cities, the influence is relatively slight. TABLE 4. — MORTALITY FROM CANCER IN CITIES, ACCORDING TO SIZE, No. of Cities 14 67 49 Size 1,000,000 and over 250,000 to 1,000,000 Less than 250,000 PERIOD 190 Population 1912 30,872,254 31,907,716 6,740,333 69,520,303 3-1912 Aggregate Population 147,889,255 148,806,139 31,840.816 Deaths from Cancer 137,531 133,286 28,064 298,881 Rate per 100,000 Population 93.0 89.6 88.1 130 328,536,210 91.0 The cancer mortality statistics for the United States are limited to the registration area, which in 1913 comprehended about 65 per cent, of the total population. Table 5 shows first the estimated mortality from cancer in the continental United States, and second the actual mortality as reported for the registration area. The rates returned for the latter have been applied to the population estimates of the former TABLE 5.— ESTIMATED MORTALITY FROM CANCER IN CONTINENTAL UNITED STATES AND ACTUAL MORTALITY FROM CANCER IN UNITED STATES REGISTRATION AREA, PERIOD 1900-1913 Est. No. Population Cancer Death Rate Deaths Population Deaths Rate per Continental per 100,000, U. S. Registration Area from United States from 100,000 Year United States Cancer Reg. Area Cancer Population 1900 75,994,575 62.9 47,829 30,794,273 19,381 62.9 1901 77,592,344 64.3 49,890 31,370,952 20,171 64.3 1902 79,190,113 65.1 51,542 32,029,815 20,847 65.1 1903 80,787,882 68.3 55,153 32,701,083 22,325 68.3 1904 82,385,651 70.2 57,794 33,349,137 23,395 70.2 1905 83,983,420 71.4 59,931 34,094,605 24,330 71.4 1906 85,581,189 69.1 59,155 41,983,419 29,020 69.1 1907 87,178,958 70.9 61,840 43,016,990 30,514 70.9 1908 88,776,727 71.5 63,494 46,789,913 33,465 71.5 1909 90,374,496 73.8 66,731 50,870,518 37,562 73.8 1910 91,972,266 76.2 70,099 53,843,896 41,039 76.2 1911 93,570,036 74.3 69,494 59,275,977 44,024 74.3 1912 95,167,806 77.0 73,282 60,427,247 46,531 77.0 1913 96,765,576 78.9 76,319 63,298,718 49,928 78.9 For the year 1914 it would seem an entirely safe assumption that the aggregate mortality from cancer approaches, if it does not exceed, 79,000. CANCER STATISTICS— UNITED STATES 7 The comparative mortality of males and females from cancer in the United States registration area for the period 1900-1913, is shown in Table 6. TABLE 6.— MORTALITY FROM CANCER, ALL ORGANS AND PARTS, BY SEX, UNITED STATES REGISTRATION AREA, PERIOD 1900-1913 r -Males Deaths Rate per r —Females — Deaths ^ Rate per Year Population from 100,000 Population from 100,000 Cancer Population Cancer Population 1900 15,415,757 7,294 47.3 15,378,516 12,087 78.6 1901 15,742,434 7,706 49.0 15,628,518 12,465 79.8 1902 16,111,848 7,798 48.4 15,917,967 13,049 82.0 1903 16,489,113 8,422 51.1 16,211,970 13,903 85.8 1904 16,856,270 8,881 52.7 16,492,867 14,514 88.0 1905 17,274,352 9,189 53.2 16,820,253 15,141 90.0 1906 21,322,133 11,166 52.4 20,661,286 17,854 86.4 1907 21,899,144 11,800 53.9 21,117,846 18,714 88.6 1908 23,876,529 13,046 54.6 22,913,384 20,419 89.1 1909 26,020,431 14,913 57.3 24,850,087 22,644 91.1 1910 27,606,526 16,373 59.3 26,237,370 24,666 94.0 1911 30,463,411 17,525 57.5 28,812,566 26,499 92.0 1912 31,128,193 18,464 59.3 29,298,940 28,067 95.8 1913 32,681,358 20,045 61.3 30,617,806 29,883 97.6 During the year 1913 it is shown that the cancer death rate for males was 61.3 and the rate for females was 97.6. There was, there- fore, an actual excess in the mortality of women of 36.3 per 100,000 of population; or, in other words, to every 100 deaths from cancer among women, there were only sixty-three deaths from cancer among men. The cancer mortality returns for most countries are on the basis of the international classification of causes of death, reduced to seven groups, respectively, buccal cavity, stomach and liver, peritoneum and intestines and rectum, female generative organs, female breast, skin, and other or not specified organs. On account of their special impor- tance, the statistics for the registration area are given for certain selected groups in Tables 7, 8 and 9. TABLE 7.— MORTALITY FROM CANCER OF THE STOMACH AND LIVER, BY SEX, UNITED STATES REGISTRATION AREA, PERIOD 1900-1913 ^-Males and Females — ^ r -Males ^ females N Deaths Rate per Deaths Rate per Deaths Rate per from 100,000 from 100,000 from 100,000 Cancer Population Cancer Population Cancer Population 1900 6,918 22.5 3,418 22.2 3,500 22.8 1901 7,095 22.6 3,594 22.8 3,501 22.4 1902 7,483 23.4 3,681 22.8 3,802 23.9 1903 8,193 25.1 4,037 24.5 4,156 25.6 1904 8,744 26.2 4,340 25.7 4,404 26.7 1905 8,939 26.2 4,388 25.4 4,551 27.1 1906 10,946 26.1 5,443 25.5 5,503 26.6 1907 11,596 27.0 5,779 26.4 5,817 27.5 1908 13,044 27.9 6,537 27.4 6,507 28.4 1909 14,915 29.3 7,477 28.7 7,438 29.9 1910 16,475 30.6 8,135 29.5 8,340 31.8 1911 17,365 29.3 8,698 28.6 8,667 30.1 1912 18.517 30.6 9,215 29.6 9,302 31.7 1913 19,767 31.2 9,749 29.8 10,018 32.7 CANCER STATISTICS— HOFFMAN TABLE 8.— MORTALITY FROM CANCER OF THE FEMALE GENERATIVE ORGANS AND FEMALE BREAST, UNITED STATES REGISTRATION AREA, PERIOD 1900-1913 Year 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 -Female Generative Rate per iths 100,000 >m Total icer Population 96 8.8 19 9.3 33 9.5 89 10.1 36 10.3 37 10.7 90 9.7 88 10.2 50 11.2 14 11.2 47 11.4 07 11.3 89 11.7 06 12.2 c -Female Breast \ Rate per Rate per Deaths 100,000 100,000 from Total Female Cancer Population Population 1,400 4.5 9.1 1,621 5.2 10.4 1,734 5.4 10.9 1,777 5.4 11.0 2,019 6.1 12.2 1,994 5.8 11.9 2,421 5.8 11.7 2,590 6.0 12.3 3,023 6.5 13.2 3.585 7.0 14.4 3,730 6.9 14.2 4,190 7.1 14.5 4,356 7.2 14.9 4,514 7.1 14.7 TABLE 9.— MORTALITY FROM CANCER OF THE SKIN, BY SEX, UNITED STATES REGISTRATION AREA, PERIOD 1900-1913 Year 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 —Males and Deaths from Cancer 602 683 688 752 758 818 984 1,121 1,282 1,492 1,459 1,619 1,743 1,725 Females—^ Rate per 100,000 Population 2.0 2.2 2.1 2.3 2.3 2.4. 2.3 2.6 2.7 2.9 2.7 2.7 2.9 2.7 t " L11CS ■>, Deaths Rate per from 100,000 Cancer Population 392 2.5 456 2.9 454 2.8 484 2.9 462 2.7 539 3.1 656 3.1 724 3.3 827 3.5 988 3.8 952 3.4 1,011 3.3 1,079 3.5 1,128 3.5 r ■ -re 111 cllCS ^ Deaths Rate per from 100,000 Cancer Population 210 1.4 227 1.5 234 1.5 268 1.7 296 1.8 279 1.7 328 1.6 397 1.9 455 2.0 504 2.0 507 1.9 608 2.1 664 2.3 597 1.9 Cancer is largely, if not primarily, a function of age, or more accurately, of senility and presenility. The mortality at ages under 35 is relatively small, and is chiefly due to sarcoma, for which, however, no separate returns are required under the international classification. Table 10 exhibits the mortality from cancer by sex and age in the United States registration area, for the period 1903-1912. TABLE 10.— MORTALITY FROM CANCER OF ALL ORGANS OR PARTS, BY AGE AND SEX, UNITED STATES REGISTRATION AREA, PERIOD 1903-1912 , Males s , Females N , Increase , Ages Deaths Rate Deaths Rate or at from per 100,000 from per 100,000 Decrease Death Cancer Population Cancer Population Actual Per cent Under 10 1,170 2.5 984 2.2 — 0.3 12.0 10-24 2,028 3.1 1,844 2.8 — 0.3 9.7 25-34 3,757 9.0 7,891 20.6 + H.6 128.9 35-44 10,750 32.3 26,779 89.0 + 56.7 175.5 45-54 24,431 105.4 46,669 222.9 + 117.5 111.5 55-64 35,327 257.4 50,393 386.4 + 129.0 50.1 65-74 33,745 452.8 43,010 565.7 + 112.9 24.9 75 and over 18,381 620.2 24,601 734.1 + 113.9 18.4 All ages * 129,784 55.7 202,421 90.6 + 34.9 62.7 45 and over 111,884 236.5 164,673 366.4 + 129.9 54.9 Including unknown ages. CANCER STATISTICS— BY AGES 9 It is shown by this table that, excepting ages under 25, the cancer death rate of females actually and relatively exceeds the rate of males, but the relative excess is highest at ages 35 to 44, when the rate for females exceeds the rate for males by 175.5 per cent. The question whether the observed increase in cancer has affected all ages or only certain periods of life is naturally of great importance. In Tables 11 and 12 the cancer death rate for males and females for the United States registration area are compared for the two periods, 1903-1907 and 1908-1912. TABLE 11.— MORTALITY FROM CANCER OF ALL ORGANS OR PARTS, BY AGES, AMONG MALES, UNITED STATES REGISTRATION AREA, PERIODS 1903-1907 AND 1908-1912 , 1903-1907 , , 1908-1912 , , Increase ^ Ages Deaths Rate Deaths Rate or at from per 100,000 from per 100,000 Decrease Death Cancer Population Cancer Population Actual Per cent. Under 10 455 2.4 715 2.6 + 0.2 -f- 8.3 10-24 840 3.2 1,188 3.1 — 0.1 —3.1 25-34 1,527 9.0 2,230 9.0 35-44 4,497 33.4 6,253 31.7 — 1.7 —5.1 45-54 9,509 103.8 14,922 106.5 + 2.7 + 2.6 55-64 13,463 245.7 21,864 265.1 + 19.4 + 7.9 65-74 12,588 427.4 21,157 469.5 + 42.1 + 9.9 75 and over 6,466 553.9 11,915 663.2 + 109.3 + 19.7 All ages * 49,458 52.7 80,326 57.7 + 5.0 + 9.5 45 and over 42,026 224.1 69,858 244.6 + 20.5 + 9.1 * Including unknown age: TABLE 12.— MORTALITY FROM CANCER OF ALL ORGANS OR PARTS, BY AGES, AMONG FEMALES, UNITED STATES REGISTRATION AREA, PERIODS 1903-1907 AND 1908-1912 , 1903-1907 » , 1908-1912 , , Increase > Ages Deaths Rate Deaths Rate or at from per 100,000 from per 100,000 Decrease Death Cancer Population Cancer Population Actual Per cent Under 10 381 2.1 603 2.3 + 0.2 + 9.5 10-24 753 2.8 1,091 2.9 + 0.1 + 3.6 25-34 3,302 20.7 4,589 20.4 — 0.3 —1.4 35-44 11,068 90.2 15,711 88.1 — 2.1 —2.3 45-54 18,788 223.0 27,881 222.7 — 0.3 —0.1 55-64 20,196 378.1 30,197 392.1 + 14.0 + 3.7 65-74 16,559 540.9 26,451 582.5 + 41.6 + 7-7 75 and over 8,916 667.0 15,685 778.6 + 111.6 + 16.7 All ages * 80,126 S.7.7 122,295 92.6 + 4.8 + 5.5 45 and over 64,459 354.9 100,214 374.3 + 19.4 + 5.5 * Including unknown ages The results of the comparison are extremely interesting and of exceptional practical utility. The increase in cancer mortality has naturally varied considerably according to the organs and parts of the body affected. It would carry me too far to discuss these important aspects of the cancer prob- lem in detail, but for the purpose of emphasizing the importance of special cancer studies, Tables 13 and 14, showing occurrence of cancer of the female generative organs and of the female breast, are included. 10 CANCER STATISTICS— HOFFMAN TABLE 13. — MORTALITY FROM CANCER OF THE FEMALE GENERATIVE ORGANS, UNITED STATES REGISTRATION AREA, PERIODS 1903-1907 AND 1908-1912 Ages , 1903-1907 » Deaths Rate , 1908-1912 s Deaths Rate ,. Increase i or at from per 100,000 from per 100,000 Decrease Death Cancer Population Cancer Population Actual Per cent. Under 10 8 0.0 23 0.1 + 0.1 + 125.0 10-24 134 0.5 236 0.6 + 0.1 + 20.0 25-34 1,158 7.3 1,831 8.2 + 0.9 + 12.3 35-44 3,836 31.3 5,984 33.6 + 2.3 + 7.3 45-54 5,810 69.0 9,090 72.6 + 3.6 -f- 5.2 55-64 4,529 84.8 7,391 96.0 + 11.2 + 13.2 65-74 2,440 79.7 4,463 98.3 + 18.6 + 23.3 75 and over 886 66.3 1,870 92.8 + 26.5 + 40.0 All ages * 18,840 20.6 30,907 23.4 -f- 2.8 + 13.6 45 and over 13,665 75.3 22,814 85.2 + 9.9 + 13.1 * Including unknown ages. TABLE 14.— MORTALITY FROM CANCER OF THE FEMALE BREAST, UNITED STATES REGISTRATION AREA, PERIODS 1903-1907 AND 1908-1912 Ages , 1903-1907 , Deaths Rate , 1908-1912 s Deaths Rate , Ii lcrease ., or at from per 100,000 from per 100,000 D ^crease Death Cancer Population Cancer Population Actual Per cent. Under 10 8 0.0 10-24 15 0.1 34 0.1 25-34 343 2.2 575 2.6 + 0.4 + 18.2 35-44 1,683 13.7 2,900 16.3 + 2.6 + 19.0 45-54 2,667 31.6 4,861 38.8 + 7.2 -f- 22.8 55-64 2,684 50.3 4,362 56.6 + 6.3 + 12.5 65-74 2,061 67.3 3,622 79.8 + 12.5 + 18.6 75 and over 1,330 99.5 2,506 124.4 + 24.9 + 25.0 All ages * 10,801 11.8 18,884 14.3 + 2.5 + 21.2 45 and over 8,745 48.1 13,351 57.3 + 9.2 + 19.1 Including unknown ages. The estimated total mortality from cancer by organs and parts of the body in the continental United States for the year 1913, on the basis of an aggregate mortality of 73,279, is shown in Table 15. TABLE 15.— ESTIMATED TOTAL MORTALITY FROM CANCER, BY ORGANS AND PARTS, IN CONTINENTAL UNITED STATES, 1913. Rate Number Percentage per 100,000 of of Population Deaths Distribution Buccal cavity 3.11 3,007 3.94 Stomach and liver 31.23 30,215 39.59 Peritoneum, intestines and rectum 10.47 10,128 13.27 Female generative organs 12.17 11,776 15.43 Female breast 7.25 7,021 9.20 Skin 2.73 2,633 3.45 Other organs or parts 11.92 11,539 15.12 All organs and parts 78.88 76,319 100.00 Important variations are found in the cancer death rates according to organs and parts of the body, and for the purpose of illustration, the cancer mortality for the cities of Boston and San Francisco is shown in Tables 16 and 17. CANCER STATISTICS— BY LOCALITY 11 TABLE 16.— MORTALITY FROM CANCER IN BOSTON, MASS., BY ORGANS AND PARTS, ACCORDING TO SEX, 1903-1912 , — Male and Female — ,, , Male ^ , Female ,, Deaths Rate per Deaths Rate per Deaths Rate per from 100,000 from 100,000 from 100,000 Organ or Part Cancer Population Cancer Population Cancer Population Buccal cavity 30S 4.9 248 8.0 60 1.9 Stomach and liver 2,027 31.9 918 29.5 1,109 34.2 Peritoneum, intestines and rectum. 1,127 17.7 446 14.3 681 21.0 Female generative organs 921 14.5 921 28.4 Breast 657 10.3 7 0.2 650 20.1 Skin 82 1.3 44 1.4 38 1.2 Other or not specified organs 1,318 20.7 679 21.8 639 19.7 All organs and parts 6,440 101.3 2,342 75.2 4,098 126.5 TABLE 17.— MORTALITY FROM CANCER IN SAN FRANCISCO, CAL., BY ORGANS AND PARTS, ACCORDING TO SEX, FROM JULY 1, 1906, TO JUNE 30, 1913 r- Male and Female—^ , Male N q Female ^ Deaths Rate per Deaths Rate per Deaths Rate per from 100,000 from 100,000 from 100,000 Organ or Part Cancer Population Cancer Population Cancer Population Buccal cavity 186 6.5 172 10.6 14 1.1 Stomach and liver 1,377 48.0 878 54.1 499 40.1 Peritoneum, intestines and rectum. 442 15.4 223 13.7 219 17.6 Female generative organs 406 14.2 406 32.6 Breast 253 8.8 1 0.1 252 20.2 Skin 67 2.3 41 2.5 26 2.1 Other or not specified organs 467 16.4 336 20.8 131 10.6 All organs and parts 3,198 111.6 1,651 101.8 1,547 124.3 Aside from locality, the cancer death rate is materially modified by race. Table 18 shows the rate for the District of Columbia for the decade ending 1 1910. TABLE 18.— MORTALITY FROM CANCER IN THE DISTRICT OF COLUMBIA, U. S. A., 1901-1910, BY AGE, SEX AND RACE; RATE PER 100,000 OF POPULATION Ages Under 10 10-19 20-29 30-39 40-49 50-59 60-69 70 and over All ages 40 and over Males Females 1.7 0.6 4.2 1.7 5.8 3.1 23.2 56.0 62.5 162.2 182.4 347.3 413.7 456.4 610.6 556.9 70.6 104.8 217.1 312.0 Males Females 2.7 1.1 9.7 13.1 26.3 72.3 48.7 207.3 139.6 328.9 310.1 386.6 335.1 522.1 38.6 86.5 130.2 293.9 It is shown that at all ages of 40 and over, the cancer death rate of white men was 217.1, and of colored men, 130.2. The rates, however, approach each other much more closely in the case of women, it being shown that the cancer death rate for white women was 312 and for colored women 293.9. As an interesting sidelight on the influence of race on the cancer death rate, a table for the island of Ceylon is included (Table 19), showing the cancer mortality for the different racial elements of the island for the period 1908-1912. 12 CANCER STATISTICS— HOFFMAN TABLE 19.— MORTALITY FROM CANCER BY RACES, CEYLON, 1908-1912 Deaths Rate Race Total from per 100,000 Population Cancer of Population Europeans 37,646 6 15.9 Burghers 131,252 34 25.9 Sinhalese 13,245,206 962 7.3 Tamils 5,488,143 296 5.4 Moors 1,298,270 84 6.5 Malays 63,082 2 3.2 Others 85,466 7 8.2 Total 20,349,065 1,391 6.8 The very low rate of 6.8 for the island of Ceylon is almost entirely due to cancer of the buccal cavity, which is directly attributable to the universal habit of betel-nut chewing. Granting that the cancer returns for Ceylon are not of a high degree of intrinsic trustworthiness comparable with the United States registration area or European countries, there can be no doubt but that, in the main, malignant disease is comparatively rare, not only in Ceylon, but also in India and other parts of Asia. That the observed variations in cancer frequency throughout the world are not primarily determined by possible errors of diagnosis or defective methods of death registration, is clearly shown by the comparative study of cancer death rates in civilized countries according to organs and parts of the body affected. This aspect of the cancer problem has thus far received inadequate consideration, but the data are unquestionably of the greatest practical significance in the statis- tical study of cancer occurrence. Table 20 exhibits the comparative frequency of cancer of the stomach, liver and oesophagus in thirteen different countries of the world for the period 1906-1910. TABLE 20.— COMPARATIVE FREQUENCY OF CANCER OF THE STOMACH, LIVER AND OESOPHAGUS IN THIRTEEN DIFFERENT COUNTRIES OF THE WORLD, PERIOD 1906-1910; RATE PER 100,000 OF POPULATION Switzerland 70.4 England and Wales 31.4 Holland 62.2 Ireland 31.0 Norway 61.4 United States Registration Area 28.3 Bavaria 59.4 Australia Commonwealth t 27.4 Japan* 40.0 Italy 26.2 Scotland 36.0 Cuba f 12.7 Uruguay 35.6 * Period 1909-1910. f Period 1908-1912. The international contrast presented by this table is of unusual interest. It is shown that the mortality from cancer of the stomach, liver and oesophagus was relatively higher in Uruguay and Japan, than in the United States registration area and England and Wales. No defects in registration or errors in diagnosis could possibly account for such profound differences in the specific incidence of cancer occurrence CANCER STATISTICS— VARIOUS COUNTRIES 13 limited to a thoroughly well-understood group of malignant affections. Table 20 also brings out the fact that the rate for this group of cancers was over twice as high in Switzerland and Holland as in the United States registration area. It would seem absurd to maintain in the face of this evidence that the observed differences in the rates of cancer frequency are primarily determined by accuracy in diagnosis or completeness in methods of death registration and analysis. In other words, the excess or deficiency in the specific cancer death rates by organs and parts of the body must be accounted for by determinable variations in local conditions, particularly as regards the mode of life, the food, the nutrition, etc., of the localities or countries considered. A similar comparison for eleven countries is made of the mortality of cancer of the skin in Table 21. TABLE 21.— COMPARATIVE FREQUENCY OF CANCER OF THE SKIN IN ELEVEN DIFFERENT COUNTRIES OF THE WORLD, PERIOD 1906-1910; RATE PER 100,000 OF POPULATION United States Registration Area 2.7 Scotland 1.7 Ireland 2.7 Holland 1.4 Australian Commonwealth* 2.3 Uruguay 1.1 England and Wales 2.1 Bavaria 0.8 Cuba* 2.0 Japan f 0.7 Switzerland 1.9 * Period 1908-1912. t Period 1909-1910. It is shown that Scotland, Switzerland and Japan have distinctly low rates of mortality in cancer of the skin, whereas the rates are dis- tinctly excessive for the United States registration area, Ireland and the Australian Commonwealth. Even more interesting in this respect are the wide variations in the comparative frequency rates of cancer of the female generative organs. Table 22 shows the rates for thirteen different countries of the world. TABLE 22.— COMPARATIVE FREOUENCY OF CANCER OF THE FEMALE GEN- ERATIVE ORGANS IN THIRTEEN DIFFERENT COUNTRIES OF THE WORLD, PERIOD 1906-1910; RATE PER 100,000 OF FEMALE POPULATION England and Wales 24.2 Italy 16.0 United States Registration Area 22.1 Australian Commonwealth f 15.5 Bavaria 21.6 Holland 13.2 Switzerland 21.4 Ireland 12.8 Japan* 20.9 Uruguay 12.2 Scotland 20.6 Norway 11.5 Cuba t 18.9 ♦Period 1909-1910. t Period 1908-1912. The highest rates are shown to prevail in England and Wales, fol- lowed by the United States registration area and Bavaria. The lowest 14 CANCER STATISTICS— HOFFMAN rates are for Norway, Uruguay and Ireland. If the argument were sound that a low cancer death rate must be considered evidence of imperfect diagnostic skill or defective methods of death registration, Norway would rank first as regards diagnosis of cancer of the stomach, liver and oesophagus and last as regards diagnosis of cancer of the female generative organs. It would also follow that since the rate for Scotland was about the same as for Japan, the diagnosis of cancer of. the female generative organs was about equally well-developed in these two countries, though widely at variance on the basis of the same 1 hypothesis in cancer of the skin. Finally, the comparative frequency rates for cancer of the female breast are of special interest, since this form of cancer is, perhaps, the most accurately diagnosed in the entire group of malignant diseases. Table 23 shows the mortality from cancer of the female breast in thirteen different countries of the world per 100,000 of female population. TABLE 23.— COMPARATIVE FREQUENCY OF CANCER OF THE FEMALE BREAST IN THIRTEEN DIFFERENT COUNTRIES OF THE WORLD, PERIOD 1906-1910; RATE PER 100,000 OF FEMALE POPULATION England and Wales 17.9 Bavaria 9.1 Scotland 15.4 Norway 7.3 Ireland 14.0 Italy 5.2 Switzerland 13.6 Cuba* 4.5 United States Registration Area 13.3 Uruguay 3.7 Australian Commonwealth* 10.6 Japan f 1.8 Holland 9.6 ♦Period 1908-1912. t Period 1909-1910. ! Recalling that the mortality from cancer of the female generative organs was practically the same in Japan and Scotland, it is shown by Table 23 that there was an immense disparity in the relative fre- quency rates for cancer of the female breast; in fact, the rate for England and Wales is almost exactly ten times the rate for Japan. Such differences as these are conclusive evidence that the variations observed in crude cancer death rates are evidently determined by pro- found differences in the specific cancer frequency according to organs and parts of the body affected. It is in this direction that statistical research can be of most value to the cancer cause, for by determin- ing the precise differences in local variations of incidence according tc the seat of primary growth, it may be possible in time to ascertain thej contributory conditions or circumstances responsible for these observed and clearly established variations. The attainment of this purpose! however, requires the perfection of the rules of statistical practice the adoption of standard methods of tabulation and analysis, the gen' CANCER STATISTICS— TENDENCY 15 eral use of the international classification of causes of death, and the extended use of the details of that classification by specified organs or parts of the seat of primary growth. These brief considerations are merely intended to emphasize the practical utility of statistical research in the furtherance of the cancer cause. The subject is of vast extent, extremely complicated, but fruitful of useful results. " Among important lines of collateral statistical research a brief reference may be made to the study of precise correlation of the com- parative frequency or changes in frequency occurrence of cancer and Dther more or less allied diseases, such as benign tumors, biliary zalculi, etc. Furthermore, this line of inquiry should be extended to nclude such diseases as diabetes, appendicitis, rheumatism, gout, syphilis, tuberculosis, malaria, etc. Most of the present-day conclu- sions regarding the correlation of cancer to any and all of these, as veil as other, diseases are based on a rather superficial consideration )f all the statistical evidence which requires to be taken into account. 50 far as practicable, extended consideration has been given to these nore involved aspects of the cancer problem in my forthcoming work >n The Mortality from Cancer Throughout the World. In its final analysis the statistical method is of the first order of mportance in determining the tendency of the cancer death rate. In trict conformity to the law of large numbers, the conclusions improve p accuracy in proportion to the amount of data considered. Table 24 [hows the increase in the cancer death rate in the New England States, New York and New Jersey, combined, for the period 1886-1913. ABLE 24. — MORTALITY FROM CANCER IN THE NEW ENGLAND STATES, NEW YORK AND NEW JERSEY, PERIOD 1886-1913 Deaths Rate per Years Population from 100,000 Index Cancer Population Number 1886-1890 55,320,449 26,215 47.4 100.0 1891-1895 64,879,439 34,536 53.2 112.2 1896-1900 71,405,669 44,645 62.5 131.9 1901-1905 78,132,762 55,501 71.0 149.8 1906-1910 87,343,060 69,140 79.2 167.1 1911 18,699,051 15,980 85.5 180.4 1912 18,976,968 16,640 87.7 185.0 1913 19,327,233 17,385 90.0 189.9 It is shown that the relative cancer death rate, assuming the rate Dr the period 1886-1890 as 100, is now 189.9, or, in other words, there as been an increase in the cancer death rate of 89.9 per cent, during te intervening period of twenty-three years. A similar comparison is presented in Table 25, for twenty large ,..merican cities for the period 1881-1913. 16 CANCER STATISTICS— HOFFMAN TABLE 25. — MORTALITY FROM CANCER IN TWENTY LARGE AMERICA* CITIES, PERIOD 1881-1913 Deaths Rate per Years Population from 100,000 Index Cancer Population Number 1881-1885 30,328,347 14,735 48.6 95.9 1886-1890 35,302,944 17,884 50.7 100.0 1891-1895 40,912,510 22,513 55.0 108.5 1896-1900 47,016,267 28,533 60.7 119.7 1901-1905 53,386,935 37,727 69.5 137.1 1906-1910 60,116,913 47,701 79.3 156.4 1911 12,849,687 10,713 83.4 164.5 1912 13,125,121 11,203 85.4 168.4 1913 13,400,553 11,971 89.3 176.1 Assuming, again, the period 1886-1890 to be represented by 100, th relative rate for 1913 was 176.1, in other words, in the twenty citid referred to, during the last twenty-three years the cancer death ratj has increased 76.1 per cent. In conclusion, the comparative rates are shown for twenty Ameri can and ten large European cities, in Table 26 : TABLE 26.— COMPARISON AS TO MORTALITY FROM CANCER, IN TWENT AMERICAN AND TEN EUROPEAN CITIES; PERIOD 1881-1912 , American Cities ^ , European Cit es ^ Cancer Death-Rate Cancer Death-Rate Years per 100,000 Index per 100,000 Index Population Number Population Number 1881-1885 48.6 100 75.4 100 1886-1890 50.7 104 82.0 109 1891-1895 55.0 113 87.9 117 1896-1900 60.7 125 97.2 129 1901-1905 69.5 143 106.2 141 1906-1910 79.3 163 114.4 152 1911 83.4 172 114.7 152 1912 85.4 176 118.3 157 Differen in Rate 26.8 31.3 32.9 36.5 36.7 35.1 31.3 32.9 The evidence presented by these tables, as well as the addition; information derived from other equally trustworthy sources, uncoi ditionally confirms the conclusion that cancer is relatively on tl increase throughout the civilized world, and that the increase is affec ing practically all important organs and parts of the body, but chief for the age period of 50 and over. It would therefore seem entire appropriate to refer to the present situation as a menace to civilize mankind, and to insist on the duty of the heartiest and most thorough nation-wide cooperation in the effort to bring about an aroused publ interest in a subject, than which, considering the enormous amount human suffering and loss of life, none can appeal more powerfully the sympathy and scientific interest of the world. COLUMBIA UNIVERSITY LIBRARIES This book is due on the date indicated below, or at the expiration of a definite period after the date of borrowing, as provided by the library rules or by special arrangement with the Librarian in charge. DATE BORROWED DATE DUE DATE BORROWED DATE DUE ^ 1/3 hr 1 / ' 1 • ?-'■- -- V C28(842)MSO C niiiffiiftK^? ITY libraries H6718 :*;•■': f