Public health reports

              PUBLIC HEALTH REPORTS
VOL. 47 FEBRUARY 5, 1932 NO. 6
THE INCIDENCE OF EPIDEMIC INFLUENZA, 1918-19*
A FURTHER ANALYSIS ACCORDING TO AGE, SEX, AND COLOR OF THE RECORDS OF
MORBIDITY AND MORTALITY OBTAINED IN SURVEYS OF 12 LOCALITIES
By ROLLO H . BRITTEN, Senior Statistician, United States Public Health Service
Introduction .
Total epidemic morbidity (influenza
incidence) .
Frequency of pneumonia as a com-plication.
CONTENTS
Introduction
Mortality and case fatality .
Summary.
Acknowledgments.
This report, one of a series of papers from this office on the epi-demiology
of influenza, is devoted to a discussion of the incidence of
the disease in the different sex, age, and color groups of the popula-tion
during the 1918-19 epidemic, as indicated by surveys made at
that time by the United States Public Health Service in certain
localities. Summaries of the results obtained in these surveys were
published shortly after the surveys were finished ; I but, as a further
contribution to the epidemiological studies of the disease, it seems
desirable to give a more detailed account of the results at this time.
The Public Health Service conducted special surveys in a number
of widely scattered localities as soon as the 1918-19 epidemic in these
' From the Office of Statistical Investigations, in cooperation with the Office of Industrial Hygiene
and Sanitation, United States Public Health Service.
I Influenza in Maryland: Preliminary Statistics of Certain Localities . By W. H. Frost and Edgar
8ydenstricker. Pub. Health Rep., Mar. 14, 1919. Reprint No . 510.
Epidemiology of Influenza. By W. H. Frost and Edgar Sydenatricker. Pub. Health Rep., Aug. 15,
1919. Reprint No . 550. (Reprinted from J. A.M.A ., vol. 73, No. 5, Aug. 2, 1919.)
Statistics of Influenza Morbidity, with Special Reference to Certain Factors in Case Incidence and
Case Fatality. By W. 11 . Frost. Pub. Health Rep., Mar. 12, 192a Reprint No. 589.
Variations in Case Fatality during the Influenza Epidemic of 1918. By Edgar Sydenstricker . Pub.
Health Rep., Sept . 9, 1921 . Reprint No. 992.
A list of epidemiologieal studies of influenza made by the Public Health Service will be found at the
end of this article.
95893°-32-1 (303)
February 5. 1932 204
places appeared to have reached its close. The purpose was to
determine for a population of known sex, age, and color composition
the approximate incidence of the disease in sample areas of a number
of widely scattered localities, and also to determine the relations
between cases of influenza, cases of pneumonia, and deaths from
these causes in so far as the number of observations would permit.
It was necessary to limit the surveys for the most part to localities
in which the Public Health Service was at the time maintaining
previously established organizations prepared to collect the requisite
data reliably and efficiently ; but in so far as practicable, the com-munities
were chosen to represent the different geographical sections
of the United States . Reference to Table 1 will show that, with the
exception of the far West, this object was accomplished in a reasonably
satisfactory manner. San Francisco was the only city west of San
Antonio, Tex ., and Des Moines, Iowa.
The survey included (a) 10 cities, varying in population from
22,500 to 680,000 ; (b) certain small towns of Maryland ; and (c) one
rural county of Maryland . The minor towns surveyed in Maryland
are usually treated as a single statistical group in this report . In the
case of Charles County, the entire population, rather than a sample
of it, was made the basis of the survey. This particular survey was
made by employees of the U. S. Bureau of the Census, funds having
been transferred to that bureau by the Public Health Service for the
purpose. The data were tabulated and analyzed by the Public
Health Service . Although the canvass included the whole county,
one of 12 enumeration districts was later dropped from the records,
owing to the presence of a proving ground (Indianhead) which
made that district unrepresentative of a general population.
In the case of Louisville, the canvass was made before the wave
of the epidemic had run its full course ; but in all the other localities
the canvass is believed to have comprised practically the whole of
the epidemic period. In Baltimore and San Francisco second surveys
were made in January and February, respectively, to obtain a record
of recrudescences which had taken place in the interval . The cases
occurring during these recrudescences are included in the data here
reported .
In the case of Spartanburg, S. C ., some time after the completion
of the canvass in the city itself, an additional survey was made of
adjacent mill villages . These villages had a disproportionately large
population of one selected class-mill workers-and for this reason the
Spartanburg data are not altogether comparable with those collected
in other localities.
The canvasses were made as soon as possible after the subsidence
of the autumn (1918) wave of the epidemic in each locality . The
following table will show the dates on which the surveys were begun
and ended :
305 February 5, 1932
TABLr 1.-Localities in which 1918-19 surveys were made, with dates of surveys,
estimated total populations, and number of persons canvassed
e Estimated as of July 1, 1918 ; revised on the basis of other data.
s The population included in survey made in November and December was recanvassed in January
in e order to record cases occurring during a recrudescence of the epidemic . Total number of persons canvassed in minor Maryland towns was 12,482.
4 Rural area .
e Census as of Mar. 12, 1919.
e Actual count in February-March, 1919 .
7 One enumeration district was later excluded from the study (see p. 304), leaving data for 16,147 can-vassed
persons. e The population included in survey made in November and December was recanvassed in February
in order to record cases occurring during a recrudescence of the epidemic.
The population estimates contained in the third column of Table
1 require some comment . Since the epidemic occurred while this
country was at war, a number of factors (principally the withdrawal of
males for military service) tend to make population estimates more
than usually unreliable in the present instance.2 By the time of the
1920 census the unusual distribution had given way to a more normal
one. An estimate based on the 1910 and the 1920 censuses will thus
not afford a reliable indication of the population of individual locali-ties
in the fall of 1918. Indeed, a satisfactory estimate is impossible,
however it be derived . But since the data here presented deal almost
entirely with actually enumerated populations in sample areas, esti-mates
of the total population are employed in only a few instances .
The estimates adopted for use in the table are based on a number of
factors, including an intercensal estimate of the population (calculated
arithmetically), allowance having been made for the withdrawal of
males for military service ; population estimates based on the normal
death rates from all causes, exclusive of respiratory infections ; infor-e
This question has been given detailed consideration in the article, "Difficulties in Computing Civil
Death Ratesfor 1918", by Edgar Sydbastricker and Mary L. King. Public Health Reports, Feb. 13,1920.
Reprint No. 58,4.
Dates of canvass
Total
Population
canvassed
Locality popula-tion
(esti- Per cent
Begun Completed mated) 4 Number of total
of persons popula.
tion
New London, Conn----------------------- Dec. 2, 1918 Dec. 18,1918 25,000 7,933 31.7
Baltimore, Md---------------------------- Nov. 20,1918 Jan. 31,19198 680,000 33,361 4.9
Minor Maryland towns: a
Cumberland--------------------------- Dec. 1, 1918 Dec. 6,1918 27,300 5,194 19.0
Frederick------------------------- .---- Nov. 27, 1918 Nov. 30,1918 11,340 2,311 20.4
Lonaconing---------------------------- Dec. 4,1918 Dec. 11,1918 2,000 1,730 86.5
Salisbury------------------------------ Dec. 10,1918 Dec. 12,1918 9,000 1,727 19.2
Downsville 4--------------------------- Dec. 7,1918 -----do-------- 850 718 84.4
LinganoreDistrict (Frederick Co .)4-__ Nov. 29,1918 -----do .------- 1,000 698 68.8
Quantico4----------------------------- Dec. 1,1918 Dec. 10, 1918 2,000 114 5.7
Charles County, Ald---------------------- (') (5) 4 18,326 718,326 100.0
Spartanburg, S. C------------------------- Dec. 5, 1918 Dec. 31,1918 22, 500 5,257 23.4
Augusta, Oa------------------------------- Feb. 1, 1919 Feb. 8,1919 55,000 4,123 7.5
At aeon, Oa-------------------------------- Dec. 4,1918 Dec. 14,1918 50,000 7,905 15.8
Des Moines, Iowa ------------------------- Jan. 31,1919 Feb. 8, 1919 115,000 5.857 5.1
Louisville, Ky----------------------------- Dec. 6,1918 Dec. 27, 1918 245,000 12,002 4.9
Little Rock, Ark-------------------------- Dec. 2, 1918 Jan. 13,1919 65 .000 9.920 15.3
San Antonio, Tex-------------------------- Dec. 5,1918 Dec. 22,1918 150,000 12,534 8.4
San Francisco. Calif----------------------- ___-_do-------- Feb. 21,19198 475,0001 18,682 3.9
February 5 . 1932 306
mation secured by Public Health Service officers located in the indi-vidual
localities ; and other available information .
Data were collected by intelligent inspectors working under specific
instructions and careful supervision . In each locality these inspectors
made a house-to-house canvass in 10 or more enumeration districts
so situated geographically as to give, presumably, a fair sample of the
general population of the city. Each district contained approximately
the same number of families . Homes at which information was not
available when the inspector called (owing to the absence of the adults,
or for other reasons) were not counted. The effort was made to
canvass in each city not less than 5,000 persons, in order to give a group
sufficient for simple statistical analyses, and in cities of more than
100,000 population to increase this number so as to give not less than
5 per cent of the total population. These conditions were generally
fulfilled .
Regarding each individual in the canvassed populations, the inspec-tors
recorded the name, color, sex, and age at last birthday ; whether
or not sick since September 1, 1918, with " influenza," " pneumonia,"
or illness suspected to be influenza (classed as "doubtful") ; date of
onset, duration, and severity of such illnesses (whether " severe,"
" moderate," or " light ") ; and date of death, if death resulted. Re-garding
each household, the inspectors recorded the number of rooms
occupied, and their impressions of the economic status of the family
(whether "well-to-do," "moderate," "poor," or "very poor") . This
point was recorded by the inspectors without instructions as to the
possible definitions of each class.'
In making inquiry as to the type or nature of illness, the enumerators
were instructed to ask which members of a family had "influenza,"
" flu," " grippe," " pneumonia," or "colds" since September 1, 1918.
Persons who were said to have been only " feeling badly," or as having
a "cold" were recorded as "doubtful" cases . If, however, the illness
lasted not less than three days and was of such severity as to confine
the patient to bed for the whole of one day, the case was classed as
"influenza," unless otherwise diagnosed by the attending physician .
Cases of illness, if definitely stated to be due to some cause other than
" influenza,"" pneumonia," or " colds," were not recorded . In view of
the difficulties of diagnosis of influenza and the large number of mild
cases indistinguishable from common colds, it was believed that the
total morbidity from influenza during the epidemic period could be
best represented by a figure which would include cases classified during
the canvass as " influenza," " grippe," " pneumonia," and " doubtful ."
The widespread nature of the epidemic minimized the effect of minor
I A special study of the data secured in relation to economic conditions has recently been issued : The
Incidence of Influenza Among Persons of Different Economic Status during the Epidemic of 1918. By
Edgar Sydenstricker. Pub. Health Rep., Jan . 23,1931, vol. 48, No. 4, (Reprint No . 1444 .)
We
Sept . 7
Sept . 1
Sept . 2
Sept . 2
Oct . 5_
Oct . 12
Oct . 19
Oct . 26
Nov . 2
807 February 5, 1932
respiratory illnesses unassociated with influenza. The inclusion of
"pneumonia" in the figures was, of course, logical, since during the
epidemic only a comparatively few pneumonia cases occurred which
were not sequelae of influenza.'
The sources of error involved in the method of survey outlined are
fully appreciated. Although the canvasses were made as soon as
possible after current morbidity and mortality reports indicated that
the wave of the epidemic had subsided, certain important points had
been forgotten by the informants . Especially was this true in regard
to the dates involved .
Another source of error arose from the fact that the families' state-ments
were accepted as to diagnosis for a disease the diagnosis of
which is especially difficult and uncertain. I\o other course was open ;
and it is confidently believed that, owing to the peculiar and wide-spread
nature of the epidemic, the data obtained were sufficiently
reliable when used in the mass.
A third source of error lay in the employment of enumerators not
specially trained for this work. However, they were carefully selected
and the inquiries were purposely made sufficiently simple to permit
even untrained persons to obtain the data with such detailed written
instructions as were furnished, if under careful supervision.
When due allowance is made for the inevitable errors incident to
the method employed, it is still believed that the surveys gave data
which represented with reasonable accuracy the influenza morbidity
in the localities surveyed. This view is corroborated by a comparison
of the chronological incidence of influenza cases in the surveyed popu-lations
and the chronological reported mortality for the population as
a whole . In the following table this comparison is made for those
surveyed localities for which death rates for the total populations
were available by weeks .
That the inclusion of "doubtful" cases was justifiable for the epidemic period, for the purposes to
which the data were to be put, is clearly indicated in the following table, from which it will be seen that,
in Baltimore (the largest sample canvassed), cases classified as "influenza," "pneumonia," and "doubt-ful"
show almost identical chronology . It is to be observed that the "doubtful" cases represent only
11 per cent of the total epidemic morbidity in Baltimore ; for the surveys as a whole such cases were 7 per
cent of the total-3,216 out of 42,920 :
Cases reported by ' Cases reported by
informant as- Cases informant as- Cases
elassftfed classified
'k ended- as I WWeeeekk ended-
"Influ- "doubt-
I~
" " " Pneu- doubt
, . grippe monfa"
tnl" f enza,"
"grippe
Influ-"
monia "
ful l,
28 6 7 Nov . 9------------- 86 12 18
4___ ._,------ 52 2 8 Nov . 16____________ 47 7 15
1------------ 126 10 14 I
Nov . 23____________ 29 6 16
8____________ 271 32 41 Nov . 30 ------------ 24 3 15
_____________
_ _ __
1,363
1,6
135
137
165
170 Tctal____________ 5,636 490 736
73 156 Percentage of all
----
-
--------
n5
44 ~ I cases_-_-_-_-_-_-_-_-_-_--_-_-_ 82.1 7.1 10.7
------------- I I
February 5, 1932 308
TABLE 2.-Weekly death rates per 100,000 from influenza-pneumonia in total
population and weekly influenza case rates per 1,000 in canvassed populations of
six localities, by weeks during epidemic of 1918-19 1
1 Deaths classified according to date of death; cases classified according to date of onset.
The mortality rates are seen to follow the case incidence rates with
considerable exactness, when one takes into account the necessary lag
due to the difference between date of onset of the disease and death
from it . So far as these few examples justify any conclusion it would
appear that, for comparison between communities, with respect to
chronology, mortality statistics give results quite similar to those
derived from morbidity statistics. In the section on case fatality,
however, it will be shown that entirely misleading results as to actual
incidence of the disease would be obtained from judging by mortality
alone.
Total Epidemic Morbidity (Influenza Incidence)
GENERAL ASPECTS
The observations made during the surveys relate to 146,203 persons,
42,920 cases, and .730 deaths. In view of the fact that the record of
the morbidity from influenza practically disappears between epi-demics
and is extremely incomplete during epidemics, special signifi-cance
must attach to the results of such a canvass. Although the
data can not in themselves give an accurate picture of the incidence
of the disease or of its case fatality in diverse parts of the country,
they do indicate the incidence and fatality for the samples surveyed
Baltimore I Cumberland Augusta Louisville Little Rock San Francisco
Week ended- Death Case
rate in DeathI Case
rate in irate
Death I Caso
rate in Death Case
rate in DeathI Case
rate in
IDeath Case
gyrate in rata in
tctal I can-rate
in
total can- in
total can- rate is
total can
_ rat? in rate ini
total i can
vassed~
_
total can-aticn
Pcpu_
lation lation
vassed
P'Pu- tat%n
vassed
P'PU lation
vassed
P'P' iatpcn ~oPn_ ~tPII vassed
Popu .
Ration . soon lation anon
i
lation
1918
Sept. 7--------- 0 1.2 0 2.1 0 1.5 0 1.7 1.5 1.0 1.3 1.5 S-pt.14-------- 1.0 1.9 0 3.1 0 .7 1.2 1.1 0 1.8 1.3 2.5
Sept . 21 _______ .7 4. 5 0 7. 7 0 1.7 3.7 1.5 0 4.9 2.7 3.5
S,pt .28-------- 2.8 10 3 3.7 33.5 1.8 2.9 1.6 1.4 0 8.6 3.2 3.3 Oct. 5__________ 17.2 49.8 33 .0 96.6 3.6 11 .2 5.7 22.9 18.5 87.3 2.9 7.2 Oct. 12--------- 82.8 57.3 307.7 123.2 16.4 14.6 37.6 8.6 133.8 95.8 6.3 13.2
Oct. 19--------- 199. 6 43 .0 402.9 71 .4 30 .9 14 .1 73.5 13 .9 146.2 52.1 27.4 27.5
Oct. 25 --------- 157.8 18.8 172.2 25.8 61.8 7.3 73 9 5.8 93.8 27.9 116.2 28.8
Nov. 2--------- 58.4 10.5 76.9 127 54.5 17.0 24.2 10.5 24.6 I 20.3 155.4 16.2
Nov. 9--------- 21.6 3.5 40 .3 . 6.7 32.7 11 .6 23.7 5.3 9.2 8.9 87.2 9.5
Nov. 16-------- 7.5 2.1 22.0 4.8 34.5 21 .6
14.3
15.91 9.6 7.7 9.8 41.7 9.4
Nov. 23-------- 5.3 1.5 14 .6 4.0 43.6 15.0 7.0 4.6 7.5 18.9 5.1
Nov. 30-------- 5.9 1.3 7.3 2.1 34.5 13.1 25.3 14 .3 13.8 7.9 11 .8 4.3 Dec. 7------.-__ 8.5 1.1 3.7 1.2 23.6 11 .4 22.4 18.8 12.3 &1 10.5 9.0
Dec. 14--------- 10.0 1.0 7.3 ------- 16.4 8.2 37.1 6.6 12.3 3.9 14.9 8.2
Dec. 21--------- 10.9 1.4 7.3 ------- 127 16.5 224 1.5 9.2 3.1 28.8 6.8
Dec. 28--------- 8.4 2.5 7.3 ------- 10.9 20.1 15.1 _______ 10.8 2.4 37.5 124
1919
Tan. 4---------- 7.1 2.5 3.7 __ 29.1 26.7 9.0 __ _ 10.8 __ _ 40.8 9.6
Jan. 11 --------- 11.0 4.0 _______ ------- 63.6 44 .9 8.2 ------- 13.8 __-___- 61 .1 7.1
Jan. 18--------- 122 3.3 ______________ 70.9 33 .2 8.6 _______ 36.9 _______ 65.3 6.2 Jan. 25_________ 22.1 1.3 ------- _______ 65.5 16 .0 12.2 _______ 21 .5 31 .4 1.3
Feb. 1__________ 20.3 .2 _______ _______ 25.5 5.1 8.2 _______ 20.0
_______
_______ 124 -------
309 February 5, 1932
and thus-in view of the correlation chronologically with the more gen-eral
records noted in the introduction-for the particular cities in which
the surveys were made. Accordingly, they serve as a check upon
the precision of other morbidity data, and indicate in a general way
certain highly important relations between morbidity and mortality.
The general incidence of influenza ("total epidemic morbidity") in
the areas canvassed will be the first point to be taken up. In a later
section of the report it will be shown that this incidence was not
greatly different in the white and colored population. Because of
this fact, and because of the small proportion of colored in most of the
localities, no considerable error will be introduced into the following
discussion by combining the white and colored rates .
TABLE 3.-Incidence of influenza in canvassed populations of each surveyed
locality during the epidemic of 1918-19
Survey made before epidemic had ended.
The rate for all localities is 294 per thousand persons. In other
words, one out of every three or four persons in the canvassed popu-lations
reported that they had influenza during the autumn wave of
the epidemic and the recurrence. Other studies made by the same
method in various parts of the country give substantially the same
results, and a tabulation of these studies by Jordan is of interest at
this point.
TABLE 4.-Incidence of influenza (autumn wave, 1918) in canvassed populations of
various United States communities 1
i From Epidemic Influenza, by E . O . Jordan, p . 190 .
I Sotne Statistics of Influenza in Oswego and Watertown in 1919. Official Bun . N.Y. State Department
of Health, 4:5 .3.
Report of Bureau of Local Health Administration . State Department of Health of New Jersey, 42:28 .
4 Statistics of the 1918 epidemic of influenza in Connecticut . Winslow, C .-E . A ., and Rogers, J . F . Journ .
Infect . Dis ., 26 :185.
It is of interest to contrast these results with those for the Army,
remembering that in the latter case the population is concentrated at
Rate Num- Num- I Rate Num- Num-
Locality per ber of ber of Locality per her of I ber of
1,000 cases persons 1,000 cases persons
All localities- . -------- 294 42,920 146,203 Baltimore, Md------------- 246 8,199 33,361
San Antonio, Tea__________ 535
--
6,701 12,534
DesDloines,Iowa ----------
San Francisco, Calif________
231
215
1,353
4,021
5,857
18,682
Minor Maryland towns____ 405 5, 060 12,482 Spartanburg, S. C__________ 214 1,126 5,257
Charles
County,
Md_______ 405 6,546 16,147 Macon 213 1,681 7,905
Little Rock, Ark___________ 359 3,565 9,920 New London,Conn________ 185 1,466 7,933
Augusta, Ga--------------- 341 I 1,405 4,123 Louisville, Ky. ,------------ 150 1,797 12,002
Rate I Num. I Numher I Rate Num-lof
Number
Locality per ber of of persons Locality per I ber of persons
1,000I cases canvassed 1,000 cases canvassed
Oswego, N. Y?__________ 470 6,094 12,952 Watertown, N. Y?_______ 282 5,765 20,473
11(illville, N. J.3 ---------- 406 4,749 11,686 Gloucester, N . J .3___-____ 245 2,930 11,969
BrIdgeton,N .J?__-______ 289 3,315 13,319 I Now Britain, Conn!____ 234 645 2,757
February 5, 1932 310
those ages when the incidence was particularly high. The rates for
four months of 1918 (September-December), corresponding approxi-mately
to the period covered by the Public Health Service surveys,
are given in Table 5. The rates are for hospital admissions for influ-enza,
bronchitis, broncho-pneumonia, and lobar pneumonia combined,
and are exclusive of sickness occurring among the troops in Europe .
TABLE 5.-Incidence of total respiratory diseases' in Army in the United States,
(admissions) September to December, inclusive, 1918 z
Rate per 1,000______ 310.4
Cases____424, 074
Mean strength______ 1,366,016
I Influenza, bronchitis, broncho-pneumonia, and lobar pneu ,nonia .
a Compiled from data given in the Medical Department of the United States Army in the World War.
Vol. IX . Communicable and Other Diseases. Prepared by Lieut. Col . Joseph F . Siler. Chapter 2 : Inflam-matory
Diseases of the Respiratory Tract, by Maj . Milton W. Hall.
With this picture before us, we are able to establish in a broad way
what the incidence of influenza was during the 1918 epidemic, and the
results secured in the surveys by the Public Health Service seem to
give a rather representative mean.
Detailed house-to-house surveys in England, comparable to the
canvass by the Public Health Service, were made in a number of towns
for the summer and autumn waves of 1918, giving considerably lower
rates than those indicated for this country. Table 6 Summarizes
these results (also from Jordan).
TABLE 6.-Comparison of influenza incidence rates per 1,000 in English towns
1918 1
I From Epidemic Influenza, by E . O . Jordan, p . 194 .
a Analysis of the results of a block census undertaken in Manchester in December, 1918. 1920 . Ministry
of Health . Report cn the Pandemic, of Influenza, a 1918-19 . London . P . 4.,6 . By T. Camwath . Report on an inquiry into the recent epidemic of influenza in the county borough of Leicester. 1920.
Ministry of Health . Report on the Pandemic of Influenza, 1918-19. London . P. 445 . By M. B . Arnold .
4 Report on incidence of influenza in the University and Borough of Cambridge, and in the Friends School, Saffron Walden . 1920. Ministry of Health . Report on the Pandemic of Influenza, 1918-19. Lon-don
. P .389. By S . M. C opeman .
5 Report on an investigation of the incidence and effects of influenza among the population of Warrington
(Lanes.) . 1920. Ministry of Health . Report on the Pandemic of Influenza, 1918-19 . London . P . 539.
By O . R' . N. Joseph .
e Analysis of an influenza census at Newcastle-upon-Tyne. 1920. Ministry of Health. Report on the
Pandemic of Influenza, 1918-19 . London . P. 556 . By 9. J . Clegg.
Returning again to the canvass made by the Public Health Service,
it will be noted that the highest rate was in San Antonio, where one out
of every two persons reported having the disease . The range of
variation in the rates is considerable, the rate in San Antonio being
nearly three times that in New London . The canvassed populations
are so large that only a relatively small part of this fluctuation can be
Locality Summer Autumn Total Poisons
Manchester :________________ 149 103 252 4,666
Leicester )__._____________________ 63 146 209 4,619 Cambridge 4--------------------------------------------------- 36 165 201
Warrington s________ 75 82 157
_____1
.626
tie%castle-upon -Tynea________________________________________ 62 47 109 4,461
explained as being due to chance." However, in several widely
separated localities the incidence rate varied onlywithin narrow limits.
A cursory examination of the rates in the different localities will
show that no consistent relation is manifested between the rates and
the geographic position of the localities. If the New England and
Maryland localities are grouped together and contrasted with the
central and southern localities, the rates in the two groups will be
found to be practically identical, namely, 304 and 306, respectively.
A marked selective effect on the incidence of influenza was exerted
by age during the epidemic of 1918-19. This observation, which is
common to nearly all reports on the epidemic, is corroborated by the
data secured in the surveys. What they show most clearly is a very
heavy incidence in the younger ages and a definite contrast with the
curve of mortality.
The influenza morbidity rates for each 5-year age group for all
surveyed localities are given in Table 7.
TABLE 7.-Incidence of influenza among canvasser) persons in each age group in
all surveyed localities during the epidemic of 1918-19
AGE
I Includes 566 of unknown age. 2 Includes 4,277 of unknown age.
February 5, 1932
It will be noted that the incidence was highest in the age group
5 to 9, fell off progressively in the age groups from 10 to 24, rose to
6 Even in the case ofNew London, whichhasone of the smallest surveyed populations, theprobable error
of the rate is less than 7 per 1,000 persons. This calculation is based on the formula
0.6745or (r.te) (1000-rate) n'~! n
wherepis the chance that an individual will have a case, qthe chance that be will not, and n the size of the
canvassed population . The probable error is applicable because therewere relatively few instances where
one person reported having more than one case.
Age group Rate per
1,000
Number of j
cases I
Number of
persons
All ages ------------------------------------------------------ .. 294. 42, !120 } 146, 203
Under 1______________________________________________________________ 207 586 2838
1-4___________________________________________________________________ 337 4,016 11,933
Under 5______________________________________________________________ 312 4, 6C2 14,771
5-9___________________________________________________________________ 391 5.755 14 .725
381 5,404 14,192
15-19_________________________________________________________________ 345 4,448 12,897
20-24_________________________________________________________________ 324 3.967 12.2,S7
2.5-29_________________________________________________________________ 337 4.127 12,234
30-34_________________________________________________________________ 326 3, 805 11,668
35-39_________________________________________________________________ 296 3,276 11, 074
40-44__________________ .______________________,_______________________ 236 2,219 I 9,415
4.5-49 ________________________________________________________________ 207 1. 688 8,157
50-54_________________________________________________________________ 175 1,162 6,628
55-59_________________________________________________________________ 162 698 4,393
60-fN_________________________________________________-_______________ 143 .537 3 756
65-69________________________________________________,________________ 1.3 .5 332 2,
,
456
70-74_ _______________________________________________________________ 111 189 1,703
75 aril over___________________________________________________________ 88 115 1,650
February 5. 1932 312
a minor second mode in the age group 25 to 29, and then declined
progressively in successive age groups. Among old people the
incidence appeared to be not more than one-third of that-, ong
the young. _;;,I 1"-
Through the courtesy 8f the heghh officers of the States of kansas
and Maryland, reports of cases of influenza in these States were
available for statistical analysis. Without going into the results of
these studies in any detail, a comparison by age is of interest for
corroborative purposes . There was, of course, no expectation that
any great proportion of the cases occurring would be reported to
the health departments of 'the States, but it was felt that the
relative incidence by age might not be greatly affected by this lim-itation.
In order to permit a comparison between the surveyed
data and the data for the two States, the rates have been reduced to
an index basis by dividing by the rate for all ages. Thus the three
curves are put on a relative basis, and the actual height becomes of
no significance .
TABLE 8 .-Relative incidence of influenza by age in surveyed localities, in Kansas,
and in Maryland during epidemic of 1918-19 (rate for each age group divided
by rate for all ages)
The results are represented graphically in Figure 1 . In general, the
curves for Kansas and Maryland correspond to the curve for the
survey, although the former show a tendency to fall off more rapidly
with age. This may be due to a greater tendency not to report sick-ness
among old people to the hearth authorities . At all events, it is
the similarity of the three curves, rather than any differences, which
is most striking .
The age curves in each of the surveyed localities may next be con-sidered
. These curves are given, in 5-year age groups, in Figure 2
and Table 9 . For the graph, as in the preceding case, the ratios of
the rate in each age group to that for all ages are used so that the age
incidence in the different localities may be readily compared.
Age group Surveyed
localities
Ban-sas
Mary-land
Age group Surveyed
localities
Ban-sas
Mary-land
Under 5---------------- 1 .01 0.73 0 .73 4549------------------- 0.69 0.59 0.58
5-9 -------------------- 1 .30 1 .28 1.29 50-54------------------- .58 -34 {
.40
10-14-------- _---------- 1 .27 1 .34 1 .36 &5-59------------------- .54 .31
I5-19------------------- 1 .15 1 .34 1.47 60--64------------------- .48 .24
20-24------------------- 1 .08 1.30 1 .40 65-69------------------- .45
25-29 1 .12 1 .36 1 .29 " 19
7'
--------- .37 20
30-34 1 .09 1.34 1.22 75 and over____________ .29
35- . .99 1.09 .96 All ages________________ 1 .00 1.00 1.00
4 .79 .88 .71
1 1
313 February 5, 1932
Tesi*_;9 .-Incidence of influenza by age in each locality during epidemic of 1918-19
jlq : [Rate per 1,000]
FIGURE 1 .-Relative incidence of influenza by age in surveyed localities, in Kansas, and in Mary-land,
during 1918-19 epidemic (ratio of rate in each age group to that in all ages)
Although minor differences are noted in the incidence in various
age groups, the essential similarity in the different localities-if we
neglect the actual level of the rates already considered-is much
more striking than these slight differences, indicating quite conclu-sively
that the selective incidence in relation to age was a marked
characteristic of this epidemic in each locality . The peak in the
1 .80
1 .60
MARYLAND
1 .40
\KANSAS
1 .20
Ol .oo SURVEYS
1
' . \
1
.so . v
.40
0-
10 20 30 30 50 60 70
AGE
-~i-c Z ;
Age ou,
New
Lo n-don
more
Balti-Mary-~
Mino
land
towns
C ount y,
a . charles l Spar-
Md.
~~n-t
rg
Au-gus"
to c°
I Des
Moines
Louis-ville
Little
Rock
San
AAnn--
to-nio
San
Fran"
cisco
Under 5-------------- 180 283 414 380 252 388 247 1 274 238 366 488 209
5-9---- _-------------- 230 366 493 448 253 480 318 350 268 463 609 281
10-14----------------- 224 317 512 486 238 416 264] 233 211 460 625 290
15-19_----------------- 177 289 493 508 232 325 219' 220 142 384 598 235
20-24 ----------------- 207 275 476 493 250
i
326 207 240 169 335 590 236
25-29----------------- 236 314 485 465 221 412 225 261 143 392 598 262
30-34----------------- 210 295 488 441 217 388 202 249 198 378 520 258
35-39----------------- 221 229 421 407 214 398 238 235 135 386 527 225
40-44 ----------------- 173 185 321 349 168 242 196 219 113 262 464 135
45-49 ----------------- 169 158 300 277 158 298 142 162 93 278 410 157
50-54 ----------------- 121 135 266 255 152 284 132 138 84 213 379 121
55-59 ----------------- 72 131 211 229 124 298 160 161 64 199 330 97
60-64----------------- 108 124 183 211 130 247 89 140 66 222 234 8o
65-69----------------- 103 112 201 181 75 j 283 63 125 83 136 294 72
70-74----------------- 74 79 145 147 132 ! 214 67 138 51 211 247 78
75 and over----------- 20 56 109 119 150 275 45 36 49 236 230 33
February 5. 1983 314
younger ages, with a gradual decline in the rates after age 30 or 35,
is found in every locality .
Perhaps of greatest interest is the suggestion that the double peak
indicated in the data for all surveyed localities and in the reported
morbidity for Kansas is really significant . The only curve which
Fiauas 2.-Ratio o[ influenza case incidence in each age group to that in all ages in a canvassed
population of each surveyed locality
does not give a suggestion of the two peaks is that for a rural area
(Charles County) . The first peak usually occurs in the age group
5 to 9 and the second peak in the age group 25 to 29. This bimodal
tendency is analyzed in Table 10, giving the age group in which the
two modes occur in each locality.
1 .50jj
NEW LONDON BALTIMORE MINOR MARYLAND
Tr TOWNS
1.25 /~
1 .00"- _.-
.75 I
.501
W .25~ a 0~
JJ
. .
Q 1.25
-Z L00_
a .75
t
.50
.25 .
o uc
DES MOINES
Zu
wa 1 .50-
Z 1 .25-
w 1.00'
F
.75'I
.50 .''
O
.25i~ OH
0
LITTLE ROCK SAN ANTONIO SAN FRANCISCO
1.25 I
1 .00 ~-
.75
.50-
.25-
0 6--T, --.--, .
20 40 60 80 20 40 60 80 20 40 60 80
A G E
315
TABLE 10.-Age groups when first and Pecond modes occur in each surveyed locality
during epidemic of 1918-19
r Same rate for 0-4 and for 5-9 .
In practically every case the second mode is quite definite, but it
should be pointed out that in only one locality (New London) is the
second mode higher than the first.
Reference may be made to the fact that W. T. Vaughan, in a house-to-
house survey of 10,000 persons in Boston, also found two peaks of
age incidence .
Question arose as to the advisability of adjustment of the rates for
influenza in the various surveyed localities to a, standard age or age
and sex distribution of the population. Such adjustments were
worked out, but found to be too slight in their effect to warrant their
use in this paper, except for certain comparisons between the sexes .'
INCIDENCE OF INFLUENZA IN THE TWO SEXES
11 ebruary 6, 1932
The morbidity rate of influenza as obtained in these canvasses was
slightly higher for women than for men, the rate for all localities
being 307 and 294, respectively, after adjustment to a standard age
6 To bring out the rather slight effect of adjustment for age and sex, the following table is reproduced .
The rates for the different localities differ somewhat from those used previously, because in this case it was
necessary to base the rate on persons of known ages.
Actual
rate per
Rate per 1,000 adjusted to standard
population (all surveyed localities)
Age group when-
Locality
First mode Second
occurs mode occurs
Spartanburg_______________ 1 0-9 20-24
Ba115more------------------ 5-9 25-29
New London_______________ 5- 9 25-29
Augusta -------------------- 5- 9 25-29
Des Moines________________ 5- 9 25-29
Little Rock________________ 5-9 25-29
Louisville__________________ 5- 9 30-34
Macon ____________________ 5- 9 35-39
San
Antoni----------------
10-14 25-29
San Francisco______________ 10-14 25-29
Minor Maryland towns____ 10-14 30-34
Locality 1,000
known
ages
By
Rate
age
Ratio to
actual
By age
Rate
and sea
Ratio to
actual
All localities_______ .______ .________________ 301 1.01 300 1 .01
San Antonio, Tea_________z______________________
Minor
536 525
"
.98 5r-
1:02
97
towns ____ . .___________________
Charles
Maryland
County, Md____________________________
408
466
418
405
1 .02
1 .00
417
405 1 .00
Little Rock, Ark------------------- ._ .__________ 360 356 .99 354 .98
Augusta, Cla__ .__________________________________ 359 362 1 . 01 359 1. 00
Baltimore, Md__________________________________ 2a3 260 1 . 03 258 1.02
Des Moines, Iowa_______________________________ 232 235
!
1 .01 23.3 1.00
Spartanburg, 8 . C_______ .____ .__________________, 217 214 .99 212 .98
S^,n Francisco . Calif_____________ ._______________ 21- 219 1.01 218 1.01
&Sacnn, (3a_______, ._______________ .__________ .__ 213 216 1.01 212 1.00
New London, Coma ________ .____________________ 18, 189 I 1 .01 189 1.01
Louisville, Ky____________________.______________ 158 165 1 .04 165 1 .04
February 5, 1932 316
distribution . The rates for women were higher in nearly every
locality. The differences are brought out in Table 11 . Adjustment
seemed advisable, because of the possible effect of the withdrawal
of males for military duty. As a matter of fact, this adjustment
made little difference in the ratio between the two sexes, the 4mad-justed
rates being 304 and 292 for women and for men for all known
ages and 299 and 288 for all ages.
TABLE 11.-Incidence of influenza by sex in each surveyed locality (adjusted to
standard age distribution) during epidemic of 1918-19
When it is realized that in a large proportion of families the infor-mation
was secured from the wife, it seems possible that this slight
excess for women might be due to the fact that they were able to
remember their own cases somewhat better than the cases of other
members of the family. A tendency of this character has been
noted in other studies where the information was secured in this
manner.' Thus the only conclusion which is really justified is
that there was no marked difference in the rates of the two sexes.
In Table 12 and Figure 3 comparison is made by sex for the dif-ferent
ages.
TABLE 12-Incidence of influenza among canvassed males and females in each age
group, in all surveyed localities during epidemic of 1918-19
r The Illness Rate Among Males amt FeM ;iles . By E . Sydenstricker. Pub . Health Rep., vol . 42, No.
30, July 29, 1927 . (Reprint 1172 .)
Rate per I,ooo Number of cases Number of persons
Age group
Male 1 Female Male Female Male Female
I
All ages____________________________ 2881 299 19,742 23,169 68,684 77,495
Under 1----------------------------------- 214 199 301 284 1,407 1,427
1-4-------------------------------- _------- 348 325 2,081 1,933 6,984 5,945
Under 5___________________________________ 322 301 ( 2,3S2 2,217 7 .391 7,372
5-9 388 344 2,845 2,910 7,342 7,382
379 383 2,649 2,755 6,994 7,187
15-19-------------------------------------- 332 356 1,985 2,461 5,086 6,909
20-24 -------------------------------------- 298 343 ' 1,267 2,699 4,405 7,881
25-29------------------------------------ __ 348 344 1,624 2,503 4,A53, 7,281
3G-34 -------------------------------------- 320 331 1,723 2,b82 6,385 6,283
35-39------------------------------- . 295 296 1,638 1,638 5,546 . 5,527
40-44 -------------------------------------- 242 230 1,112 1,107 4,592 4,823
45-49------------------------- ------------- 200 215 850 888 4,250 3,907
50-54 -------------------------------------- 167 184 655 607 3,319 3,308
55-39-------------------------------------- . 157 166 j 334 363 2,130 2,192
60-64-------------------------------------- 128 157 237 300 1,848 1,904
65-59-------------------------------------- 132 138 154 178 1,170 1, 286
70-74
and_
___ 114 108 85 104 744 959
75 ove 83 92 58 87 702 948
Locality
All localities_______________
Rate
Male
294
per 1,000 !
Female l
307 l
Ratio of
female
rate to
male
1 .04
Locality
San Francisco______________
Rate
Male
213
per 1,000
I Female
222
Ratio of
female
rate to
male
1.04
San Antonio_______________ 514 530 1.03
Macon_ _____ ____ 194 229 1 1 .18 Augilsta ___ 357 3(34 1.02
Minor \Maryland towns____ 406 459 1 .13 Charles County, 403 406 1.01
Spartanburg___________ .___ 200 220 1 .10 Des Moines________________ 229 231 1.01
Baltimore__________________ 248 270 1 . 09 Louisville__________________ 166 164 .99
New London______________ 185 192 1 .04 Little Rock________________ 352 345 .98
317
Except for the youngest ages, there is a tendency for the female rates
to be higher, but, as just pointed out, the difference is slight. The
ratios of female rates to male rates for broad age groups are as follows :
Under 15 years, 0.99 ; 15 to 44, 1 .07 ; 45 to 59, 1 .07 ; and 60 and over,
1 .09. The age curves are practically identical in the two sexes, the
only difference between the two being the greater depression in the
male curve between the two modes. In fact, the female curve shows
only a bare suggestion of the second mode.
FIGURE 3.-Incidence of influenza among canvassed males and females in each age group (all sur-veyed
localities)
INCIDENCE IN COLORED POPULATION
February 5. 1932
Since a number of the cities had a considerable colored population,
it is of interest to determine whether a larger percentage of white or
colored were attacked . We are faced immediately with the difficulty
of getting as complete information from the colored as from the white
in a canvass of this character ; thus any results must be discounted.
No rates have been used for all the surveyed localities, because of the
varying proportion of colored persons in the different localities. The
rates in the eight places where there was a sufficient number of colored
to give somewhat reliable results are given in Table 13, adjustment
having been made to a standard age and sex distribution .
A50
"I00
.w FEMALE 350 .-
O 300
Oo_ MALE
250
wa
I'
w 200
Q
150-
-1O0__
50--
0 L F 10 20 30 40 50 u0 70
AGE
February 5, 1932
TABLE 13.-Incidence of influenza in white and colored canvassed populations
during the epidemic of 1918-19 (adjusted to a standard age and sex distribution)
318
1 Rates for Charles County unadjusted ; adjustment made only a slight difference in the ratios.
With the exception of Charles County, Md. (see p. 304 for informa-tion
as to method of survey in this locality), the rates are consistently
lower for the colored populations . In Louisville, Baltimore, and
Augusta the rate is at least twice as great in the white as in the colored
population . The fact that the colored population live generally under
conditions presumably more favorable to the spread of contact infec-tions
would lead one to expect a higher rate of influenza among them.
How much of the difference is to be ascribed to more complete report-ing
among the white populations is quite impossible to determine .
Some confirmation of this difference between the incidence of influenza
in white and colored is given by the rates for the Army while in the
United States. The period covered in the table is September-
December, 1918 .
TABLE 14-Incidence of total respiratory' disease by color in Army in the United
States, September-December, inclusive, 1918 a
I Influenza, bronchitis, broucho-pneumonia, lobar pneumonia.
VI Compiled from data given in the Medical Department of the United Vo1 . States Army in the World War . 1x . Communicable and Other Diseases . Prepared by Lieut . Col . Jcsepb F. Siler. Chap. 2: Inflam- matory Diseases of the Respiratory Tract, by Maj . Milton W. Hall .
One further table is presented giving the incidence of influenza by
color in the two sexes. The tendency for higher rates in the white
population is evidently present in both sexes.
White Colored
Rate per 1,000________ 316 260
Number of cases_ _ ___ 383,498 40,576
"Strength"__________ 1,215,447 150,569
Rate per 1,000 Number of cases Number of persons
Locality
White Colored
Ratio of
colored
to white White Colored white Colored
Louisville______________________ 179 49 0.27 1,739 58 10,534 1,465
Baltimore______________________ 278 116 .42 7,600 481 29,085 4,195
Augusta________________________ 456 212 .47 1,044 361 2,434 1,689
Macon________________________- 220 137 . 62 1,337 341 5,971 1 .930
Spartanburg-___________________ 224 173 . 77 1,033 84 4,652 581
Minor Maryland towns________ 419 385 .92 4,744 249 11,782 643
Little Rock_____________________ 360 338 .94 2,657 908 7,282 2,654
Charles County, Md___________ 379 431 1 .14 3,028 3,518 7,992 8,155
319
TABLE 15.Incidence of influenza by sex and color in certain canvassed localities
during epidemic of 1918-19
95893°-32-2
GENERAL ASPECTS
Rate per 1,000
The Frequency of Pneumonia as a Complication
February 5,1932
The 1918-19 epidemic of influenza was notably different from the
1889-90 epidemic in a much higher frequency of pneumonia and
consequently a much higher mortality, especially among young
adults. The record of pneumonia cases in the areas canvassed by
the Public Health Service is therefore of interest, particularly in
view of the inadequacy of pneumonia morbidity reports during either
epidemic or normal periods . As noted in the introduction, cases
were classified in these surveys as "pneumonia" when so reported
by the householder. No attempt could be made to diagnose the
cases or to inquire of the physician in charge as to the diagnosis made
by him. Deaths from influenza were classed as pneumonia cases
even when not so specified on the census report .
The results obtained in Charles County are evidently not com-parable
to those obtained in the other localities, since in this county
there were only 102 pneumonia cases recorded, whereas there were
147 deaths from influenza-pneumonia . The deaths in this instance
were presumably complete, as the results of the survey were checked
up with the death certificates in the State registrar's office ; but since
it may be assumed that epidemic deaths were due almost always to
complicating pneumonia, and since by no means all of the pneumonia
cases resulted in death,' clearly the pneumonia cases were not com-plete.
Because of these obvious inconsistencies, the records from
Charles County have been omitted from all discussions of pneumonia
morbidity .
The following table gives the pneumonia incidence for all localities
(except Charles County) and for each locality.
" If we were to assume completeness of recording nonfatal cases of pneumonia, we would have a fatality
rate to Charles County of 82 per cent, whereas in the other localities the average is about 25 per cent .
Locality Male Female
White Colored White Colored
Louisville_________________-_________________._____r____________ 169 43 162 37
Baltimore______________________________________-_______________ 255 98 272 129
Augusta_______________________________________________________ 427 197 430 225
Spartanburg___________________________________________________ 207 135 235 152
Macon --------------------------------------------------------- 198 171 246 180
Little Rock____________________________________________________ 377 308 355 371
Minor Maryland towns ___________________________________ .____ 397 330 415 432
Charles County, Md___________________________________________ 383 419 374 445
February 5, 1932 320
TABLE 16 .-Incidence of pneumonia in canvassed population of each surveyed
locality during epidemic of 1918-19
I Exclusive of Charles County, Md.
The pneumonia case rate for all localities (except Charles County)
was 17.6 per 1,000 persons, as compared with 280, the influenza rate,
for the same localities. In other words, the percentage of influenza
cases complicated by pneumonia, as determined in these surveys, was
6.3. A more detailed comparison with influenza morbidity will be
taken up later. At this point it is desirable to summarize the pneu-monia
data themselves.
The most striking feature of the pneumonia rates is their wide
range. The minor Maryland towns have a rate four times as great
as that of Spartanburg (surveyed population, 5,257) and nearly three
times as high as Louisville (surveyed population, 12,002) .
Another point of interest is that the cities with the lowest rates
are invariably in the south central part of the country, where, it is
believed, the epidemic was somewhat less severe. The combined
pneumonia rate for Augusta, Macon, Louisville, and Spartanburg
was 10.7, whereas it teas 19 .6 in the other localities combined.
The toll of the epidemic in young adult life is depicted clearly by
the rates for cases of pneumonia recorded in these surveys . The
pneumonia incidence in each age group for all localities is presented
in Table 17. The numbers are evidently sufficient for quite reliable
results.
TABLE 17 .-Incidence of pneumonia by age in all localities, exclusive of Charles
County, Md., during epidemic of 1918-19
AGE
Rate Num- Number Rate Num- Number
Locality per be, of of Locality per her of of
1,000 cases persons 1,000 cases persons
All localities l------ 17 .6 2,290 130,056 NewLondon_____________ 17 .1 136 7,933
Minor Maryland towns__ 25 .8 322 12,482 Little Rock______________
l Augusta-----------------
16 .0
15 .3
159
63
9,920
4,123
San Antonio_____________ 24 .2 303 12,534 i Macon------------------- 13 .0 103 7,905
Des Moines______________ 23 .6 138 5,857 I Louisville________________ 9.2 111 12,002
Baltimore________________ 18 .0 599 33,361 nartanburg_____________ 6.7 35 6,257
San Francisco__ 17 .2 321 18,682
Age group Rate ner
1,000
Number
of cases
All ages________ 17.6 2,290
Under 1______________ 24 .9 60
1-4___________________ 2& 0 264
tinder 5-------------- 25 .8 324
-------------------- 14 .8 186
10-14_________________ 11 .5 137
15-19_________________ 15.5 173
20-24________-________ 23.1 256
25-29_________________ 31 .1 352
30-34_________________ 25,7 279
35-39_________________ 21 .0 213
40-44_________________ 13.0 112
45-49_________________ 9.8 73
50-59_________________ 8.3 82
60-69__
70 an d over
_
_
_
_
9.3
6.5
51
19
There are two marked peaks. The incidence is high in children
under 5 years of age, although not any higher in the first year of life
than in the years immediately following . The second mode occurs in
young adult life, the highest point being found in the age group 25 to
29, where the rate is three times that in the age group 10 to 14. As
age advances, the rate falls off rapidly. By 50 years of age it is
already one-half of the rate for the age group 25 to 29. A direct
comparison with the incidence of influenza as a whole is postponed
until later, but it may be pointed out that the bimodal effect noted
in the case of influenza is much more marked in the case of pneumonia
alone. In both the incidence falls off steadily with age after the second
peak.
So striking is this bimodal tendency for pneumonia curves according
to age during the epidemic that it seems well to present the rates by
age for the individual localities . The numbers are limited, and it has
been necessary to combine certain age groups . The data are given in
Table 18 and Figure 4.
TABLE 18.-Incidence of pneumonia in each canvassed locality, by age, during
epidemic of 1918-19 1
Age group
Rate per 1,000 persons canvassed
321 rebruary 5, 1932
I Inclusion of deaths from influenza as pneumonia cases was not possible in this table, except where the
case was originally recorded as pneumonia. The rates, however, are not more than about 7 percent too low.
The marked bimodal effect is noted in each locality without
any exception. In all but one city the first peak comes in the under
5-year age group. Usually the second peak is in the age group 25
to 29, but in three instances it is in the age group 30 to 34, and in
one in the age group 35 to 39. It is evident that the location of these
modes is subject to a certain chance variation.
This strikingly high incidence of pneumonia in the young adult
population, reaching a peak of nearly 5 per cent in some of the locali-ties
in the modal age group, is obviously at great variance with the
normal age distribution of pneumonia. An idea of this difference
may be obtained from a comparison of the age curve secured in this
canvass with that for Hagerstown, Md., during a period (December
1, 1921, to April 1, 1924) without major epidemic waves, the data
New
London
Baiti.
more
Minor
Mary-land
towns
Spar-tan-burg
Augus-ta
Macon Des
Moines
Louis-ville
Little
Rock
San
An-tonio
San
Fran-cisco
11 . 1 27.3 38. 2 10.4 29.6 17. 5 37 .4 22.2 16.3 20.4 20.2
9. 0 13 .4 21. 0 3. 2 23 . 1 11. 7 50.0 8.5 7.8 14.7 8. 7
7. 8 11 .3 15. 2 3.8 7 . 4 7. 3 14.2 9.2 10.9 11 .2 10.8
12.4 18. 5 19.3 4.3 13 . 2 10. 4 18 .0 6.1 11 . 7 20. 1 15. 2
25 .4 21 . 1 37.0 6.0 18 . 0 11 .2 25.7 9.4 24 .1 39.8 14. 7
44 .4 29. 4 39.7 9.1 25 . 9 14.8 38. 7 11.0 22,1 42 2 30. 7
28 .7 21. 8 46.2 8 .8 2& 7 14.6 24.7 11.7 24 .7 34.4 22 8
19.0 18.4 38.8 4.9l 7.4 15.7
16
.3 10.6 18 . 3 31 .9 17.7
10.4 10.7 14.9 & 1J 16.5 10. 4 4.7 9 . 8 13.7 17.7
6.9
7.2 9.31 1.2 9.2 8.3 &8l 3.6 ( 9 . 5 17.4 10.8
514 9.4 3. 99 4.5 21 8. 99 &9 8. 7 9. 0
February 5.1982 322
having been secured in house-to-house canvasses during this period
by the Public Health Service .' No comparison of the actual level
of the morbidity rates seems feasible or of consequence in this con-nection,
in view of the varying periods for which the sickness data
Ftct.TRz 4 .-Ratio of pneumonia case incidence in each age group to that in all ages in a canvassed
population of etch surveyed locality
in the various localities were secured . Comparison may be made
most easily by reducing each series of rates to an index basis by
dividing by the rate for all ages. These indices are given in Figure 5
and Table 19.
The Incidence of Various Di,eases according to Age. Hagerstown bforbidity Studies No. VIII. By
Edgar Sydenstricker. Public Health Reports, May 11, 1921+ . ( ;= g , ri=,t No . 1227 .)
ALL LOCALITIES NEW LONDON BALTIMORE
250
2.00-
1 .50
1 .00 - -
w .50
u
Q
J 0
Q 2.00 MINOTROMWANRSYLAND SPARTANBURG AUGUSTA
2T
1 .50
1 .00
1
O
.
I-w
.50='
ua
I
=
G 250-
w
MACON DES MOINES LOUISVILLE
I
200
Wa
1.50
O I .00 ..\ i
O r a .50j
. LITTLE ROCK SAN ANTONIO SAN FRANCISCO
L50
I .0 0
.so-
0!=
~"
I
0 20 40 60 0 20 40 60 0 20 40 60
AGE
FwURE 5 .-Relative incidence of pneumonia by age in surveyed localities and in Hagerstown,
aid . (data for Hagerstown from a previous sickness survey)
TABLE 19-Relative incidence of pneumonia by age in surveyed localities during
1918-19 epidemic and in Hagerstown sickness study (rate for all ages-1 .00)
323
Annual rates .
February 5, 1932
3 .00 ( To 4.60
2.80 I
V) 2 .60
I
I
V ~
Q 2 .40--
J I
2 .20
Q 2 .00
I
=l- I
1 .80-- I
F- I
v 1.60- I
a I
1 .40 . I a~
II ~w
1 .20
J
1 .00
1
/
. //
.80 \ /
O /
O \ /
.60-- \ i
.40-- \\
PGEF
t5V0
.20
10 20 30 40 50 60
AGE
Age group
Rates per 1,000
Surveyed Hegers-localities'
town
i
Indices
Surveyed Hagers-localities
town
Under 5________________________________________________________ 25.8 40 .0 1 .47 4.60
6-0 ____________________________________________________________ 1 .09
10.14
14.8
11 .5
j!
9 .6
" 7.0
.84
.66 .84
15-19___________________________________________________________ 15.5 2.0 s
.88 .23
20-24___________________________________________________________ 23.1 1 .31
2,5-29_____________________________________________________ 3l . 1 117
3(1-34_____________________________________________________ _5 . 7 3.0 1 . 46 .34
35-39___________________________________________________________ 21 .0 1 .19
411--14___________________________________________________________ 13.0 .74
45-54_':_________________________________________________________ 8.7 5.4 3
.49 .62
6
;5-64______-----------------------------------------------------
_ 9.5 . 64
65 andover_______________ 7 .9 9.9 .45 1 .14
All ages________________________________________________________ 17 .6 8.7 1.00 1 .00
February 5. 1932 324
In a nonepidemic period, pneumonia has its highest frequency at
the beginning and end of life . In the pandemic of 1918 pneumonia
showed its highest frequency in the age group 25 to 29, a subordinate
peak in the age group under 5 years, and a relatively low incidence
after 40 years of age. It should be observed that the contrast is
really somewhat greater than that shown in the figure, since the
curve for the epidemic contains a proportion of deaths from pneumonia
not associated with the epidemic and therefore tending to follow
the age curve as typified by the Hagerstown data.
In contradistinction to the material presented for the total mor-bidity
during the epidemic, the pneumonia rates are slightly higher
in the males, as shown in Table 20. Spartanburg is omitted, because
only 35 cases were recorded in all, but is included in the total for
all localities. The rates have been adjusted to a standard age
distribution .
TABLE 20 .Incidence of pneumonia by sex in each surveyed locality during epidemic
of 1918-19 1 (adjusted to standard age distribution)
SEX
In only one locality is the rate for females definitely higher. The
fact that we do not find higher rates among females for these serious
cases suggests that possibly the difference in the incidence of influenza
as a whole was due to the tendency of the women to report a higher
incidence for themselves than for other members of the family. That
would hardly be expected in the case of illnesses severe enough to be
classed as pneumonia, as they would probably be recalled whatever
member of the family had the case.
A graph is added for pneumonia incidence by sex and age . (Fig. 6 .)
There is a suggestion that the excess among males occurs entirely
during the ages where the epidemic exerted its greatest effect. The
rates are presented in Table 21 .
Locality
Augusta_____________________________________________
Rate pcr
Male I
16 .5
1,000
Female
17. 9
Ratio of
female
rate to
male
1
02
OH
Cases
?Yale I
26
Female
35
Baltimore___________________________________________ 18 .7 1P. 1 I. . 267 327
KewLondon________________________________________ 16 .7 16 . J 1.01 64 72
Macon______________________________________________ 12 .9 12 .5 .97 44 56
l.ouisville___________________________________________ 11 .2 9.6 .86 52 53
Minor Maryland towns_____________________________ 30.6 25 .1 .82 159 163
Little Rock__________________________________________ 17 .0 13 .7 .81 81 77
San Antonio_________________________________________ 27 .0 21 .7 .80 139 161
San Francisco__________.____________________________ 20.8 14 .8 .71 177 144
Des Moines_________________________________________ 29.4 19 .5 .66 79 59
1 Spartanburg omitted because of small numbers.
325 February 5, 1932
TABLE 21 .Incidence of pneumonia by sex and age in all surveyed localities during
epidemic of 1918-191 (rate per 1,000)
I Exclusive of Charles County.
OOO
W
0-
WQ
45-
40
35
30
25
20
i
15
10
5
0--
A
MALE
i--r-!-'r-
50 60 70
FiGCRS 6.-Incidence of pneumonia by age and sex in all surveyed localities (except Charles
County, Md.)
In view of these differences, it is of interest to compare the rates
by sex and age in each locality . To do so, however, a broad grouping
of ages is necessary to secure any degree of regul4rity. These broad
groups have been chosen to bring out, as well as possible, the char-acteristics
of the age curve (under 5 years, 5-19, 20-29, 30-39, 40 and
over) . The rates are given in Table 22. At the bottom of the
table will be found ratios of the female rates to those of the males.
Age group Male Female
All saes________ 18.4 17.0
Under 1______________ 28.8 21 .3
1-4___ _______________ 27.5 1.5
Under 5_ 27.7 23.8
5--Q ------------------- 14 .1 15.5
10-14 ----------------- 12.1 11 .0
15-19_-------------- __ 17 .2 14.2
20-24 ----------------- 24 .1 22.5
21-29 ----------------- 37 .4 26.9
3G-34 _________________ 29.0 22.9
35-39 ----------------- 24 .1 17.9
40-44----------------- 13 .2 12.8
45-49 ----------- 11 .0 8.6
50-59__________6.2 10.3
60-69_ 6 .1 12.2
70 and over 4 .2 8.2
February 5. 1932 326
TABLE 22 .Incidence of'pneumonnia by sex and broad age groups in each surveyed
locality during epidemic of 1918-19 1
All localities :
111
Feamlael-e------:-,
New London .
Male_ ___.
Female_
Baltimore:
Male .
Female___
Minor 3laryla
Male-----
Female ___.
Spartanhurg:
Mole_____.
Female---
Augusta :
Male_____
Female---.
Afaeon :
Male_____.
Female___
Des Moines:
:Male_____
Female___
Louisville :
Male_
Female___
Little Rock :
Male_____
Female___
San Antonio:
Male ----
Female ---
San Francisco:
Male-----
Female ---
RATES PER 1,000
RATIO OF FEMALE RATE TO MALE
COLOR
yearr5 I 5-19 I 20_29 ~ 30-39 40 and
I Inclusion of deaths from influenza as pneumonia cases was not possible in this table, except where the
case was originally recorded as pneumonia.
The tendency is toward an excess in the male rate at the ages 20
to 39 and is evidently present in a great proportion of the localities.
The recorded pneumonia incidence was generally greater among the
white than among the colored population. The following table gives
the cases and rates by color for each locality in which there was a
considerable number of colored (except Charles County) .
___ _________
.____________________________________
.____________________________________
___________________________
nd towns:
_
23.3
20.5
11.3
11 .0
30 .5
24.0
36 .2
12.2
11 .8
12.6
7.2
13.2
16.4
20 .7
26.3
223
34.7
34.1
27.9
23.3
48.4
23.8
18.2
20.9
27.3
19.4
20.6
54.5
7.6
9.3
4.9
9.5
6.2
10.8
.____________________________________ 5.9
40 .1 16.9 31.5 32.7 12.5
________________________________
---------------------------------------
3.5
17 .3
1.4
5.8
6.6
11 .3
7.3
7.0
----------
4.7
.____________________________________ 18.4 13.4 25 .4 15 .9 10 .8
.____________________________________ 40.2 15 .8 20.8 14 .3 5.8
.____________________________________
.____________________________________
20.6
14.8
7.1
11 .9
13.0
13.0
22,5
8.2
7.C
10.2
---------------------------------------
.____________________________________
49.1
2.5.4
37.5
17.7
32.4
321
21 .9
19.2
8.f
6.C
---------------------------------------
.____________________________________
19.2
25.6
9.0
&6
11 .5
9.6
15.4
7.8
3.f
4.1
_______________________________________
____________________________________
22 9
10.0
10.4
9.9
25.0
21.9
26.
16.5
7.f
9. :
______________________________________ _
____________________________________
26.7
13.8
15.3
15.5
44.6
40.6
40.
26.7
14.£
12.1
______c________________________________
_______________________________________
20.8
19.71
12.8
10 .6
31.1
17.6
24.0
16.7
19.f
10.1
All localities_________________________________________ 88 97 85 76 122
New London ---------------------------------------- 97 57 98 131 194
Baltimore------------------------------------------- 79 117 84 106 174
Minor Maryland towns_____________________________ Ill 82 65 60 212
Spartanburg---------------------------------------- _ __________ __________ __________ __________ ----------
Augusta_____________________________________________ 218 118 82 90 54
Maeon______________________________________________ 72 168 100 36 147
Des Moines ----------------------------------------- 52 47 99 88 so
Louisville_ _________________________________________ 133 73 84 51 114
Little Rock________ __________________________________ 44 95 88 63 122
San Antonio_________________________________________ 52 101 91 65 84
SanFrancisco_______________________________________ 95 83 57 70 5C
Locality
Louisville, Ky____________--____
Augusta, Ga____________________
Baltimore, Md______ .__________
Mason, Ga_____________________
Minor Maryland towns _____ .__
Little Rock, Ark_______________
Spartanburg, S . C______________
327
TABLE 23 .Incidence of pneumonia in canvassed white . and colored populations of
certain surveyed localities during epidemic of 1918-19
Mortality and Case Fatality
February 5, 1932
In some localities the colored population seemed almost to escape
the disease, while the white population was severely affected . In
Baltimore the white and colored rates were, respectively, 19.1 and 9 .3,
and in Louisville 10.1 and 2 .7 . This relation is consistent with the
fact that, in the canvassed populations, the mortality was slightly
higher in the white than in the colored .
Rates of mortality in the general population of this country during
the pandemic of 1918 have been thoroughly analyzed. There is no
occasion to refer to them in the present paper, or to utilize the record
of deaths obtained in the canvass to corroborate such findings. The
value of these records lies rather in the fact that by means of them we
may have a fairly precise conception of the case fatality of the 1918
epidemic in the communities sure-eyed . The section will deal with
the case fatality of the epidemic as a whole (the percentage which
the influenza-pneumonia deaths are of the influenza cases) and the
case fatality of pneumonia (the percentage which these deaths are
of the pneumonia cases), together with some reference to the mortality
rates themselves.
It has been previously pointed out that it is impossible to distinguish
between deaths reported as due to influenza and those reported as due
to pneumonia-in practically all cases both of these diseases contrib-uted
to the deaths. Therefore only a slight error will be introduced
in taking the relation'between the influenza-pneumonia deaths and
the total epidemic or pneumonia cases. Obviously these deaths also
include a small number of normal or nonepidemic deaths. In view
of the small size of the samples and the lack of information as to the
normal rate of pneumonia in these sample areas, it has been impossible
to limit the study to epidemic deaths alone.
In the sit communities in which a comparison was possible, it was
found that the influenza-pneumonia death rate in the canvassed popu-lation
was only about 70 per cent of that in the city as a whole during
the same period. The discrepancy was found consistently in each
Ratio of
colored
rate to
white
Number
White
of cases
Colored
Number
White
of persons
Colored
0.27 107 4 10,534 1,465
.45 48 15 2,434 1,689
.48 556 39 29,085 4,195
.80 82 21 5,971 1,930
.71 309 12 11,782 643
.80 123 36 7.262 2,654
.75 32 3 4,652 581
February 5, 1932 328
community, varying from 57 per cent in Louisville to 84 per cent in
Baltimore . The data are recorded in Table 24, which gives also the
mortality rates in the surveyed areas of the localities for which mor-tality
rates for the whole city were not determined.
TABLE 24 .-Alortality from influenza-pneumonia during epidemic period in total
populations of certain surveyed localities and in canvassed populations of same
localities
I Middle date of reconvass . 2 Exclusive of Cumberland (given above) .
There are a number of factors which may tend to explain the lower
mortality rates in the canvassed populations : (a) Deaths of nonresi-dents
in hospitals in the city have a tendency to raise the city mor-tality
rates, but would not appear in the canvassed population ;
(b) there might be a tendency for persons visited to fail to mention
deaths occurring in the family some time previously ; (c) canvassed
populations naturally do not include certain groups of the population
in which mortality rates are likely to be excessive, such as boarding
houses. Whatever the cause of this discrepancy, it is manifest that
the case fatality rates to be discussed are affected by it in some degree .
The case fatality for all localities (percentage of total cases which
were fatal) was 1 .70 . If we consider the pneumonia cases alone, it
was 25 .5 (omitting Charles County) . The data by. locality are given
in Table 25.
Death rates Deaths re- Ratio of
per 1,000 ported Death rates rate for
Middle Estimated based on from Sep- per 1,000 canvassed
Locality date of popula- reported tember 1 canvassed population
survey tion deaths in to middle to that for
total pop- date of persons total pop-ulation
survey ulation
Baltimore---------------------- Jan . 151 680,000 6 .2 4,239 5 .2 0.84
Cumberland-------------------- Dec. 3 27,300 10.8 295 7 .1 .66
Augusta ------------------------ Feb. 4 55,000 6 .3 348 4.4 .70
Louisville ---------------------- Dec. 16 245,000 3 .7 908 2.1 .57
Little Rock--------------------- Jan . 3 65,000 5.1 3 ..30 3.9 .77
San Francisco------------------ Feb . 15 , 475,000 7.8 3,700 4.8 .62
New London ------------------- Dec. 10 25,000 ------------------------ 5.8 ---------__-
Minor Maryland towns 9------- Dec. 5 26,190 ------------------------ 6.4 ------------
Charles County, Md----------- Mar. 12 18,326 ------------------------ 9.1 ------------
Spartanburg-------------------- Dec. 18 22,500 ------------ ------------ 1 .9 ------------
b1acon------------------------- Dec . 9 50,000 ------------ ------------ 3.2 ------------
Day Moines-------------------- Feb . 4 115,000
------------
3.8 ---------_--
San Antonio-------------------- Dec . 14 150,000 J------------ 4.2 -----------_
i Exclusive of Charles County, Md.
329 February 5, 1932
TABLE 25 .-Influenza and pneumonia case fatality in canvassed populations of
each surveyed locality during epidemic of 1918-19
LOCALITY
NEW LONDON
CHARLES COUNTY
SAN FRANCISCO
BALTIMORE
MINOR MD.TWNS.
DES MOINES
MACON
LOUISVILLE
AUGUSTA
LITTLE ROCK
SPARTANBURG
SAN ANTONIO
CASE FATALITY
PER 100 INFLUENZA
CASES
1 2 ..
PER 100 PNEUMONIA
CASES
1,0 2i0 30
CASES COM-PLICATED
BY
PNEUMONIA
(PER CENT)
2 4 6 8 10
T'TI-F
FIGURE 7.-Case fatality of influenza and of pneumonia, with percentage of cases complicated by
pneumonia, in specified localities
A great variation in the fatality rates is observable, which is no
doubt partly due to the small number of deaths. For total influenza,
the fatality varies from 3.14 per cent in New London to 0 .78 per cent
in San Antonio. The coefficient of variability is 37.11 The pneu-monia
fatality showed much less variation, the coefficient being 23.
The highest rate was in New London (33 .8) and' the lowest in Des
Moines (15 .9) . These fatality rates are presented by graph in
Figure 7, together with the percentage of cases complicated by
pneumonia.
10 In making this calculation the minor Maryland towns were subdivided. See p . 305. Coefficient of
variability is the standard deviation times 100 divided by the mean.
Fatality rate per P er cent Number of cases 100 cases of influ-Locality
enza com- Number plicated of deaths
Pneumo- by pneu- Pneumo-nia
monia ( Influenza nia
All localities ------------------------ 1 .70 125.5 16 .8 42,920 12,290 730
New London ----------------------------- 3.14 33.8 9.3 1,466 136 46
Charles County, \Id--------------------- 2 .25 _ ___ _ _ _____ _ _ 6,546 ------- 147
San Francisco_____________________________ 2 .24 28.0 8 .0 4,021 321
5991
90
Baltimore .-------------------------------- 2 .10 28.7 7 .3 8,199 172
Minor Maryland towns------------------- 1 .66 26.1 6.4 5,060 322 b4
Des Moines
-------------------------------
1 .63 15.9 10.2 1;353 138 ~12
Macon------------------------------------
1 .49 24.3 6.1 l,6sl 103 25
Louisville--------------------------------- 1 .39 22.5 6.2 1,797 111 25
Augusta ---------------------------------- 1 .28 28.6 4 .5 1,405 63 18
Little Rock_______________________________ 1 .09 24.5 4 .5 3,565 159 39
Spartanburg------------------------------ .89 28.6 3.1 1, 126 35 10
San Antonio ------------------------------ .78 17.2 4.5 6,701 303 52
February 5, 1932 330
Examination of the graph shows that the influenza case fatality
seemed somewhat lower in the south central part of the country.
A map has been included (fig. 8) to bring this out more clearly .
The fatality rate is indicated by symbols of varying degrees of
density .
A comparison of the influenza case fatality with that obtained in
certain other house-to-house canvasses is next given (Table 26) .
Since the available data are for the northeast section of the country,
the only rates from the Public Health Service surveys which have
been included in the table are for New London, Baltimore, and the
minor Maryland towns.
FATALITY PER 100 CASES
OFINFLUENZA
5-9
®I.0 -1 .4 .2.5 AND OVER
FIGURE 8 .-Case fatality in different cities
NEW
LONDON
MINOR MD.TOWNS
BALTIMORE
CHARLES COUNTY
TABLE 26.-Influenza case fatality rates during pandemic of 1918 in certain
house-to-house canvasses
r Statistics of the 1918 Epidemic of Influenza in Connecticut. 1920. Journ. Infec . Dis., 28 :1&5. Winslow,
C.E . A ., and Rogers, J. F .
o Some Statistics of Influenza in Oswego and Watertown in 1918-19 . Off. Bull. N. Y. Stat e Health, 4 :53 . Baker, O . Dept. of W.
Influenza : An Epidemiological Study . Am. Journ. Hyg., Monograph No. 1, 280 pp. 1921 . Vaughan, W. T .
A question arises as to whether the incidence of influenza or the
incidence of pneumonia determined tl - -rortality rates in the sur-
Locality
Case Number
of persons Locality
Case
fatality
Number
(per, of
centage) surveyed
. (per-ceutage)
persons
surveyed
II . 8. Public Health Surveys: New Britain, Conn .'-____-__ . _ 3.9 2, 757 New London_______________
I
3.1 7,933 Watertown, N. Y. I____-___ 3 .1 20,473 Baltimore________-_________ 2.1 33,361 Boston3------------------------ 2 .5 10,050
Minor Maryland towns____ 1.7
1
I 12,482 Oswego, N. Y. r---------------- 2 .4 12,952
331 February 5, 1932
veyed communities . This question can be considered from several
angles . For instance, the fact that the case fatality of pneumonia
was less variable than that of the epidemic as a whole (as previously
noted) suggests that it was the presence of the secondary invaders
which primarily determined the mortality. Another point of view
is to consider the correlation of the rates of influenza, pneumonia, and
deaths. The highest correlation is between the incidence of pneu-monia
and the mortality rates, but there is a definite correlation in the
AGE
100 200 300 400 300
INFLUENZA INCIDENCE
FIGURE 9 .-Composite picture o[ incidence and mortality In the various localities
other two instances . The coefficients are as follows : Influenza
incidence-pneumonia incidence, + .63 ; influenza incidence-mortality,
+ .66 ; pneumonia incidence-mortality, + .77. A composite picture of
the interrelations by locality is shown in Figure 9, where the height
of the vertical bars represents the mortality rates.
The age curve of mortality from the epidemic is given for all local-ities
in Table 27, first for both sexes combined and then for males and
females separately.
February 5, 1932 332
TABLE 27.-Mortality per 1,000 persons from influenza-pneumonia by sex and age
in all surveyed localities during epidemic of 1918-19
The most obvious point to be brought out is the extraordinary age
curve of mortality during the epidemic . There is no necessity of
emphasizing this fact here, since it has been thoroughly recognized
in all accounts of the 1918-19 epidemic and the contrast with the
usual experience has been apparent to everyone .
Discussion of the differences between the two sexes will be post-poned
until later. (See p . 334.)
The fatality of the epidemic according to age is of extraordinary
interest, because it brings out so clearly the severe toll among young
adults . The rates are presented in Table 28, for both the case fatality
of influenza and that of pneumonia alone.
TABLE 28 .-Fatality of influenza and of pneumonia by age, in all surveyed localities
during epidemic of 1918-19 (percentage of cases which died)
1 Exclusive of Charles County, Md.
The very high incidence of pneumonia in young adult ages (pre-viously
discussed) is evidently the most important frctor in the deter-mination
of the curves shown herewith . The fatality of influenza
rises to nearly 3 per cent in the age group 25 to 29 and then falls to
less than 1 .5 per cent. In old age it rises again, reaching 5 per cent
or more. Pneumonia cases themselves do not show this striking
change in fatality in young adult life. As a matter of fact; the
pneumonia fatality curve, except for an expected high value at the
beginning of life, rises rather consistently from 12 per cent in the age
group 5 to 9 to nearly 60 percent in old age . It must again be stressed
that the picture of pneumonia fatality includes the cases and deaths
Age group Both Male Female Age group I sees Male Female
All ages ------------ 5.0 5.3 4.1 20-24---------------------
2x29_____________________
6.2
9.9
6 .8
13.3
5 .8
7 .6
Under 1__________________ 15.2 17.1 13.3 . 30-4--------------------- 7.9 9 .1 6 .8
1-4 ----------------------- 6.2 5.4 7.1 35-39--------------------- 6.3 7 .9 4 .7
40-44--------------------- 4.0 4.1 3 .9
Under 5------------------ 7.9 7.6 8.3 45-49_ __________ 2.9 3 .5 23
5- 2.2 1.9 2.4
5i,
---------- 2.6 2.8 2 .4
10-14_ 2.1 1.4 2.8'
60-69-------------------
4.3 3 .3 5 .6
15-19 --------------------- 3.4 4.0 2.9 70 and over______________ 5.1 . 4 .2 5 .8
Age group Influenza monia- i Age group Influenza monia i
All ages________________ 1 .7 25.5
I
20-24-------------------------
25-29
1 .9
2.9
25.0
30.1
Under 1______________________
--
7 .4
-
43.3
___----------------- _----
30-34 ------------------------- 2.4 28.0
1-4 --------------------------- 1 .8 18 .6 35-39 ------------------------- 2.1 28 .6
40-44 ------------------------- 1.7 28.6
Under 5___________,__________ 2.5 23.1 45-49 ------------------------- 1.4 27.4
5-9 --------------------------- 0.6 11 .8 ! 50-59---------- 1 .5 28.0
10-14 ------------------------- 0.6 16 .1 160-69__________________________ 3.1 45.1
15-19 ------------------------- 1 .0 19.1 170 and over____ 5.1 57.9
333
nGuRF 10 .-Epidemic relations, by age, on relative basis (all ages=1 .00) . (Charles
County omitted in rates involving pneumonia incidence)
February 5, 1933
AGE
10 20 30 40 50 60 70
200
MORTALITY
180 PNEUMONIA CASE RATE
^
-=
. TOTAL MORBIDITY
160
140
120 /
100
/
.
F
80
60
w
40
JJQ
20
0
ax
1' 300
O
1-
w 280
<7 - INFLUENZA CASE FATALITY
Q -- PNEUMONIA CASE FATALITY
x 260 ---- PERCENTAGE OF CASES
U COMPLICATED BY PNEUMONIA Q
w 240
Z
220 /
200 //
O
180
F-
160
140 /
1,20
100
80--
60
40
20
0
10 20 30 40 50 60 70
A G E
February 5, 1932 334
which would have occurred at this time of year quite apart from the
epidemic.
Perhaps it would be convenient to summarize in a single graph all
the relations which have been brought out with respect to age, because
the striking manner in which the epidemic affected young adults is so
clearly depicted . Figure 10, accordingly, gives the age curves for
influenza incidence, pneumonia incidence, mortality, percentage
which the pneumonia cases were of the influenza cases, case fatality
of the epidemic as a whole and case fatality of pneumonia. The
indices (ratio of the rate in each age group to that for all ages) are
given in Table 29 .
TABLE 29 .--Ratio of rates in each age group to those in all ages in all canvassed
localities during epidemic of 1918-19
Exclusive of Charles County, Md.
SEX
As in the case of comparisons by locality, these relations indicate
that the mortality is determined primarily by the incidence of pneu-monia.
The cause of the high mortality in young adult life evidently
lies in the complicating pneumonia. All of the relations shown in
this figure bear this out : The peak in the pneumonia case incidence
in young adult life, coinciding almost completely with that of the
mortality from the epidemic ; the absence of a corresponding peak in
the total epidemic morbidity (except a minor secondary mode) and
(by corollary) a peak in young adult life for influenza case fatality
and the percentage of cases complicated by pneumonia, but not for
pneumonia case fatality itself.
Mortality and case fatility rates for influenza and for pneumo-nia
were higher among men than among women, the differences
being about 10 per cent on the average. In the case of influenza
fatality, this may have been due to the fact that the reports were
usually obtained from the female members of the household, giving
Age group incidence
Influenzal Pneumo-nia
incidence'
Per cent
compli-sated
by
pneumo-nia
i
I,
Mortality,
i
Influenza
case fatal-ity
Pnenmo-nia
mse
fatality I
All
aees------------------
:________________ 1 .00 1.00 1 .00 1.00 1.00 1.00
tinder 5___________________________________ 1 .04 1 .47 1 .37 1.59 1.47 .91
1 .42 .84 .62 .44 .35 .46
lo-14___ . 1.34 . 65 .51 .42 .35 . 63
15-1J_____________________________________ 1.18 .88 .78 .68 .59 .75
20-24_____________________________________ 1 .08 1 .31 1 .21 1 .24 1.12 .98
2:-29 . ._ ._________________________________ 1 .16 1.77 1 .51 1.98 1.71 1.18
3C-34____ . :_______________________________ 1.11 1.46 1 .29 1.58 1.41 1.10
3.~-39____________ .________________________ 1 .02 1 .19 1 .17 1.26 1.24 1.12
40-44_______________,._____________________ .80 .74 .92 .80 1.00 1.12
45--19_____________________________________ .70 .55 .78 .58 .82 1.07
50-59_____________________________________ .55 .47 .81 .52 .88 1.10
CO-00___ ---------------------------------- .42 .52 1 .11 .86 1.82 1.77
70 and over ____ .__________________________ .31 .37 1.10 1 .02 3 .00 227
a relatively higher rate of influenza among them. But a similar
explanation is hardly possible in the case of pneumonia fatality.
Table 30 gives the relations between the two sexes for all the measures
which have been employed in this report. All ratios based on pneu-monia
incidence are exclusive of Charles County, Md., as indicated .
For the other cases, all 12 localities are used. Since it was found
that adjustment for age made little difference in the ratio between
the two sexes (see p. 324), these rates are given without adjustment.
TABLE 30.-Epidemic relations by sex in all surveyed localities during epidemic of
1918-19
I Exclusive of Charles County, Md.
33,5 February 5, 1932
The mortality rates by age and sex have already been given.
(Table 27.) The excess among men would seem to occur at the ages
when the epidemic took its severest toll (20 to 40). This is equally
borne out in the fatality rates, which are given in Figure 11, especially
in the case of influenza case fatality. The two sexes evidently pre-sent
a quite different picture, which may be regarded as of importance
in connection with the epidemiological problems raised by the disease .
The data are given in Table 31 . Table 32 gives corresponding figures
for the percentage of cases complicated by pneumonia .
TABLE 31 .-Fatality of influenza and of pneumonia by age and sex in all surveyed
localities during epidemic of 1918-19
i Exclusive of Charles County.
Ratio fe-
Male Female male to
male
Influenza incidence (per 1,000)____________________________________________ 288 299 104
Pneumonia incidence I (per 1,000)_________________________________________ 18 .4 17.0 92
Percentage of influenza cases which were complicated by pneumonia i----- & 8 5. 9 87
Mortality (per 1,000)_____________________________________________________ & 3 4.7 89
Case fatality-influenza (per cent)________________________________________ 1 .8 1 .8 89
Case fatality-pneumonia alone I (per cent)_______________________________ 28 .5 24.5 92
Fatality per 100 Fatality per 100 Fatality per 100 Fatality per 100
cases of cases of caves of cases of
influenza pneumonia Age group I Age group influenza pneumonia
Males Females Males Females Males Females Males Females
All ages. .. . 1 .8 1.5 25.5 24 .5 20to24----------
29__________
2,4 1 .7 27.5 23.6
Underl_________ 8 .0 8.7 44.1 423
25 to
30to34----------
4 .1
2.8
2.2
2.1
38.3
29.9
24.5
25.9
1 to 4____________ 1 .5 22 12.9 25 .0 135 to 39__________ 2,7 1.6 31 .1 25.3
40to44---------- 1 .7 1.7 29.1 28.1
Under 5--------- 2,4 2.8 19.0 28 .0 !45to49 1 .8 1.1 28.6 25.8
6to9------------ .5 .8 11 .4 122 ~50to59__________ 1 .7 1.3 40.0 21.2
10 to 14__________ .4 .7 11 .4 20 .9 80 to 69__________ 23 3.8 43.8 45.7
15 to 19__________ 1.2 .8 221 16.1 i 70 and over______
i
4 .2 5.7 00.0 57.1
February 5. 1932 336
TABLE 32.-Percentage of influenza cases which were complicated by pneumonia, by
age and sea; in all localities, during epidemic of 1918-19 1
Exclusive of Charles County.
COLOR
Outside of Charles County, Md., the fatality rate per 100 cases of
influenza was about the same in the white and colored populations,"
1 .7 and 1 .9, respectively. The pneumonia case fatality (excluding
Charles County) in the white and colored was 28.8 and 39 .8,
respectively. Thus we are probably warranted in concluding that the
case fatality was really higher in the colored populations of the
surveyed communities.
Summary
The purpose of this report has been to make a permanent record,
for future reference, of the statistics obtained by the surveys, not to
offer any extended discussion of their meaning . Hence there is no
necessity for any detailed summary of the findings . Certain major
points, however, are of considerable interest.
Special surveys were undertaken at the close of the 1918-19
epidemic of influenza to determine for a population of known sex,
age, and color composition the approximate incidence of the disease,
and also to ascertain the relations between the epidemic morbidity,
the incidence of pneumonia, and the mortality. Preliminary reports
on the surveys were published at the completion of the work.
The incidence of influenza (including pneumonia and "doubtful"
cases) was 294 per 1,000 for all localities, varying from 535 to 150.
These rates correspond closely with what was found in other surveys
of the same general character. There seemed to be no clear indication
of a geographical difference in incidence .
The incidence was highest among very young persons (age group
5 to 9 years), with a secondary peak at about 30 years. The rate of
attack fell off rapidly in older life . Among old people the incidence
appeared to be not more than one-third of that among the young.
Slightly higher influenza rates were found among females (except
in two localities), but it seemed possible that this was due to the fact
that most of the reports as to illness came from the women, who
"New London, SanAntonio, Des Moines, and San Francisco excluded. In thecase of thesecalculations
by color, it wasnot possible to add to the pneumonia cases deaths reported as due to intluenm
Age group Both Male Female Age group sexes Male Female
sexes
All ages____________ 6.3 6.8 5.9 20 to 24___________________
25 to 29___________________
7.6
9.5
9.4
12.1
6.8
7.9
Under i------------------ 12.2 13.6 11 .0 30 to 34___________________ 8.1 9.4 7.0
1 to 4_____________________ 8.1 8.2 7.9 35 to 39___________________ 7.4 8.5 6.2
40to44------------------- 5.8 5.7 5.8
Under 5__________________ 8.8 8.9 8.3 45 to 49___________________ 4.9 5.7 4.1
5to9__________ 3.9 3.7 4.1 50to59------------------- 5.1 4.1 6.0
10 to 14___________________ 3.2 3.4 3.0 60 to 69 ___ _______________ 7.0 5.0 8.5
15 to 19___________________ 4.9 5.8 4.2 70 and over_______________ 6.9 4.5 8.5
337 February 6, 1932
might remember their own illnesses better than those of other mem-bers
of the. family. The colored had lower rates of influenza incidence,
but it is possible that the reporting among them was less complete.
A special effort was made to determine the incidence of pneumonia
as complicating the original case of influenza . For all localities the
FiGVse n .-Influenza and pneumonia fatality, by age and sex, in all surveyed localities during
the 1915-19 epidemic . (Pneumonia fatality 1s exclusive of Charles County)
pneumonia rate was 17.6 per 1,000 persons, varying from 25.8 to
6.7 . In other words, about 6 per cent of the influenza cases were
complicated by pneumonia.
The peak in young adult life suggested in the epidemic morbidity
as a whole comes out with remarkable clarity in the pneumonia
AGE
10 20 30 40 50 60 70
W
7 .0--
NQU
6 .0
N
Z
5 .0 /
Z MALE
O i
9 3 .0
iFEMALE~~~---
i 1 .0
0
W 70--
60--
Z0
50
i
40
°o MALE
30 v'`-~ /
i __-_~~
FEMALE
~ i
i ~ l0
0
10 20 30 40 50 60 70
AGE
February 5. 1932 338
incidence . For all localities the rate is about 25 per 1,000 at the
beginning of life, falls to about 11 in the age group 10 to 14, and then
rises to a secondary mode of about 31 in the age group 25 to 29.
After that the rate falls rather steadily to the end of life. This age
distribution is, of course, fundamentally different from the normal
course of pneumonia incidence, which is high among the very young
and among the very old . The striking mode in young adult life is
found in each locality without exception .
The pneumonia incidence rates were slightly higher among finales,
the difference being especially marked in young adult life. The re-corded
pneumonia incidence was higher among the white than among
the colored .
The deaths from influenza-pneumonia during the epidemic period
were obtained primarily to determine the relations as to case fatality.
The fatality per 100 cases of influenza (total epidemic morbidity)
was 1 .70 for all localities, and that per 100 cases of pneumonia alone
was 25 .5. The fatality for the surveyed localities (total epidemic
morbidity) seemed about the same as that recorded in other studies .
The fatality seemed lower in the southern and central localities,
which is in line with other reports on this epidemic . It appeared that
the incidence of pneumonia, rather than that of influenza as a whole,
determined the mortality in the various localities.
The fatality of influenza (total epidemic morbidity) was very high
among young adults, as would be expected in view of the high peak of
pneumonia at these ages. The fatality of pneumonia did not show
this peak, showing that the tendency to a severe toll at these ages was
characteristic of the pneumonia itself, rather than of death from it.
The fatality rates, both for influenza and for pneumonia, were
higher among men than among women. In the case of influenza, this
may reflect the tendency of the women to report more adequately ;
but that would hardly explain the difference in the case of pneumonia
fatality. The excess was most marked in young adult life.
The pneumonia case fatality was much higher among the colored
than among the white.
Acknowledgments
Special acknowledgment is made to the Influenza Commission of
the Metropolitan Life Insurance Co ., which defrayed part of the
expenses of the tabulation and analysis of the data collected in the
surveys.
The surveys themselves were made under the direction of Senior
Surg. W. H. Frost and Principal Statistician Edgar Sydenstricker,
and the first papers reporting the results were prepared by them.
Doctor Frost and Mr. Sydenstricker also supervised the more
extended analysis on which this paper is based.
339
Bibliography
PRECEDING PAPERS ON THE EPIDEMIOLOGY OF INFLUENZA
February 5, 1932
Preceding papers from the Office of Statistical Investigations dealing with
various phases of the epidemiology of influenza are as follows :
Age and Sex Incidence of Influenza and Pneumonia Morbidity and Mortality
in the Epidemic of 1928-29 with Comparative Data for the Epidemic of 1918-19.
(Based on surveys of families in certain localities in the United States following
the epidemics.) By Selwyn D. Collins. Pub. Health Rep., vol. 46, No. 33,
August 14, 1931 . (Reprint No. 1500 .)
The Incidence of Influenza among Persons of Different Economic Status During
the Epidemic of 1918 . By Edgar Sydenstricker . Pub. Health Rep., vol. 46,
No. 4, January 23, 1931 . (Reprint 1444 .)
Mortality from Influenza and Pneumonia in 50 Large Cities of the United
States, 1910-1929. By S. D. Collins, W. H. Frost, Mary Gover, and Edgar
Sydenstricker . Pub. Health Rep., vol. 45, No. 39, September 26, 1930 . (Reprint
1415.)
The Influenza Epidemic of 1928-29 with Comparative Data for 1918-19.
By Selwyn D. Collins. Am. Jour . Pub. Health, Vol. XX, No. 2, February, 1930.
Influenza-Pneumonia Mortality in a Group of About 95 Cities in the United
States, 1920-1929. By S. D. Collins. Pub. Health Rep., vol. 45, No. 8, Feb-ruary
21, 1930 . (Reprint 1355 .)
Morbidity in the Influenza Epidemic of 1928-29. By M. V. Veldee. Pub.
Health Rep., vol. 44, No. 19, May 10, 1929 . (Reprint 1282 .)
The Influenza Epidemic of 1926. Pub. Health Rep., vol. 41, No. 34, August 20,
1926 . (Reprint 1104 .)
Variations in Case Fatality During the Influenza Epidemic of 1918 . By Edgar
Sydenstricker. Pub. Health Rep., vol. 36, No. 36, September 9, 1920 . (Reprint
692.)
Statistics of Influenza Morbidity, with Special Reference to Certain Factors
in Case Incidence and Case Fatality . By W. H. Frost. Pub. Health Rep.,
vol. 35, No. 11, March 12, 1920 . (Reprint 586.)
The Epidemiology of Influenza. By W. H. Frost. Pub. Health Rep., vol. 34,
No. 33, August 15, 1919. (Reprint 550.)
Epidemic Influenza in Foreign Countries. By W. H. Frost and Edgar Syden
stricker . Pub. Health Rep., vol. 34, No. 25, June 20, 1919. (Reprint 537.)
Influenza in Maryland. By W. H. Frost and Edgar Sydenstricker . Pub.
Health Rep., vol. 34, No. 11, March 14, 1919 . (Reprint 510.)
A Comparison of the Mortality Rates by Weeks During the Influenza Epi-demic
of 1889-90 and During the Primary Stage of the Influenza Epidemic of
1918 in 12 Cities in the United States . Pub. Health Rep., v91.34, No. 5, January
31, 1919. (Reprint 502.)
Preliminary Statistics of the Influenza Epidemic. By Edgar Sydenstricker .
Pub. Health Rep., vol. 33, No. 52, December 27, 1918 .