Agreement of results obtained and comparison with combustion analyses.

[blocks in formation]

From the table it is evident that the agreement between the results obtained by different observers when using blood from different subjects is very good. When no CO was present, out of nine observations on three samples, only one reported it as present, and then only 0.005 per cent. The greatest percentage error reported was that in the case in which 0.000 per cent CO was recorded by one observer for one of the samples when 0.010 per cent was present. As the percentage of CO increased, the absolute error also increased, as would be expected from the nature of the dissociation curve; but in all events the results obtained are within the limits of error allowable from the standpoint of the purpose for which the method was devised.


1. A durable and compact laboratory or field apparatus for the quantitative determination of carbon monoxide in blood and air has been described. On account of the compactness, durability, and ease with which accurate and dependable results can be obtained, it should be of great use in investigations pertaining to the cause, diagnosis, and treatment of carbon-monoxide poisoning.

2. The percentage saturation of CO in blood can easily be determined to a degree of accuracy involving only 5 per cent error.

3. The actual error of the method for determination of CO in air, even when used by inexperienced men, was found to be 0.005 in regions of 0.000 to 0.05 per cent CO; 0.01 in regions of 0.05 to 0.08 per cent; 0.02 in regions of 0.08 to 0.12 per cent; and 0.03 in regions of 0.12 to 0.18 per cent carbon monoxide.


RESPECT TO EARNINGS. By FRANK M. PHILLIPS, Ph. D., and GERTRUDE A. SAGER, M. A., Office of Industrial Hygiene and Sani

tation, United States Public Health Service. During the close of 1916 and the beginning of 1917, officers of the United States Public Health Service made a physical examination of 916 men employed in the various occupations of the general chemical industry. Of this number, 629 were married men supporting one or more dependents. Data were gathered showing the monetary income of these men, making possible a classification of them by income.

The purpose of this particular study is to make comparisons by income classes, of the physical measurements, diseases, defects, and impairments as revealed by this physical examination. No attempt is made either to substantitate or to refute the conclusions of other investigators who have made similar studies. This report contains simple statements of certain conditions as they were found to exist in this particular group of persons.

The income range is not large. The 629 men examined are divided into incomes classes as follows: Incomes of $16 and under $20 per week; $20 and under $25 per week; and $25 and over per week. The numbers of men in each of these three income classes are 291, 230, and 108, respectively. The average ages of the groups are 35.07 years, 35.04, and 34.99, in the order named. Slight adjustments for age distributions are made in the rates given. No attempt is made to adjust for difference in occupation, length of service, or nationality. The majority of these men are Americans. The nationalities next in order of frequency. are Poles, then Slavs. There were a few Italians and some persons of other nationalities.

Table I shows, by income classes, certain physical measurements, certain economic and vital facts, and certain disease and defect rates per thousand for these economic groups. No attempt is made to test the reliability of the rates, but the number in each group is large enough to make the probable errors comparatively small.

TABLE 1.Certain physical measurements, diseases, and vital and economic facts con

cerning 629 married, male, white, chemical employees by income groups.

[graphic][subsumed][subsumed][ocr errors][subsumed][ocr errors][ocr errors][subsumed][subsumed][ocr errors][ocr errors][subsumed][subsumed][ocr errors][subsumed][ocr errors][ocr errors][subsumed][ocr errors]

It is shown in Table I that these groups are about the same average age and that they work about the same number of hours per day. As the occupations are nonseasonal, there is no unemployment to be charged against any group, even if difference in wages means difference in occupation.

Dynamometer readings, chest expansion, and vital capacity all increase with income. The number of children born and the number still living decrease with increase in income, whereas the per cent of children born who are still living increases with income. It thus appears that a child has a better chance of living in the better-paid group. The number of rooms per person increases, while the number of persons per bedroom decreases with increase in income. An average sized family of a little over four persons was found to have an average of nearly five rooms with an average of 1.6 bedrooms. In the lower income groups the congestion is slightly higher than this, whereas in the better-paid groups it is slightly lower.

Heart disorders and pyorrhea are both high in the lower-income class, decreasing as the income increases.

In this study, then, chest expansion, dynamometer readings, vital capacity, number of rooms per person, and per cent of children still living all show a direct correlation with income, whereas the number of children born, number of children living, number of persons per bedroom, cases of pyorrhea, and of heart disorders all show a negative correlation with income.

Blood count, as measured by the Tolquist index, shows no regular series in either income direction. The same statement may be made regarding overweight and underweight, hernia, defective vision, defective hearing, defective teeth, tuberculosis, bad posture, and diseased tonsils.

It should be remembered that results of this kind are subject to at least two interpretations of an entirely different nature unless further analysis is possible. Quite often it is the man's physical impairments and other unfortunate circumstances that rob him of both time and energy, and thus keep him out of the higher-income classes and prevent his getting a larger salary.

An observation might be made here regarding heart disorders. It is generally conceded by physicians that these occur among rich and poor alike. If so, it is possible that a bad heart condition may keep a man out of the better-paid groups by limiting his earning capacity.

On the other hand, it is also possible that lack of income renders it impossible for the employee to care for physical ills that cost money for correction.

Both of these forces operate at all times, and unless sufficient data are gathered and carefully analyzed, it is impossible to draw any definite conclusions.

There is no purpose nor advantage in this study of making any other claim than that of relationship between size of income and the items enumerated. The income range is small. The differences noted in Table I are, for the most part, small also. The statement of facts and the accompanying discussion are submitted as one more contribution to the literature on the subject of income and its relation to physical measures and bodily ills.



The question of whether or not mortality from cancer has been actually or only apparently increasing during recent years has been a matter of controversy for some time. Statisticians of the Metropolitan Life Insurance Co. are studying the cancer mortality data for wage earners insured in the industrial department of the company for the period 1911 to 1922. (There were approximately 14,000,000 persons in this group during 1922.) From a preliminary study of the rates," "standardized” to account for changes in the composition of this group of persons, with respect to color, sex, and age, the conclusion is reached that the slight upward trend in the cancer death rate during the years under consideration is due more to changes in diagnosis and recording than to any circumstances really affecting cancer incidence. It was found that the

i Statistical Bulletin, Metropolitan Life Ins. C'o., Vol. IV, No. 8, August, 1923.

2 Recorded increases in mortality from cancer of accessible sites probably represent actual increases, as diagnoses of such forms of the disease were probably no better in 1922 than in 1911.-Ed.

"standardized" or adjusted death rate for cancer increased approximately at the rate of six deaths per million policy holders per year during the period 1911-1922. The change of rate among females, for whom the cancer death rate is much higher than for males, was slight--an increase of about two deaths per million per year for white females and a small decrease, less than one death per million per year, among colored females. The death rates for males showed a higher average increase than did those for females, the increase being at the rate of about 1:1 deaths per million per year for white males and about 12 deaths per million per year among colored males. The following table is given, showing the average annual increment or decrement for this group, by color and sex, during the period under consideration:

Standardized death rates for cancer (all forms) and the uniform annual increasc? or

decrease, 1911-1922.
[Industrial Department, Metropolitan Life Insurance Co.)

[merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][ocr errors][ocr errors][ocr errors][merged small][ocr errors][ocr errors][ocr errors][ocr errors][merged small][ocr errors][ocr errors][ocr errors][ocr errors][ocr errors][merged small]

Another point brought out in this preliminary study is that the cancer death rate in this group has shown very little yearly variation from the line of general trend over the period 1911-1922, the average deviation being but 2 per cent and not over 4 per cent in any one year. This small deviation is interpreted by the statisticians of the company as signifying that no fundamental change in cancer mortality actually took place. In the following table are given the crude mortality rates for cancer among the industrial policyholders of the company for the years 1911-1922, together with the standardized death rates and the standardized trend death rates and the uniform annual increment. The crude death rates from cancer and other malignant tumors in the death registration area of the United States (exclusive of Hawaii) during the period 1911-1920 are also

« ForrigeFortsett »