1918 Analysis with Charts and Graphs

[gsgs]

Re: analysis by Frost, Sydenstricker

looks very good. How did you find it, how could I miss it ?

See also this thread:

http://www.flutrackers.com/forum/showthread.php?t=26504

[Jonesie]

Re: analysis by Frost, Sydenstricker

I FOUND IT ON THIS SITE:

http://www.history.navy.mil/library/...uenza_main.htm

[Jonesie]

Re: analysis by Frost, Sydenstricker

And if you want to start another thread on the USA Army statistics, here it is:

EXCERPTS ON

THE INFLUENZA AND PNEUMONIA PANDEMIC OF 1918

FROM

WAR DEPARTMENT ANNUAL REPORT TO THE SECRETARY OF WAR

FISCAL YEAR ENDING JUNE 30, 1919



REPORT OF THE SURGEON GENERAL, U.S. ARMY

IN TWO VOLUMES

http://history.amedd.army.mil/booksd...G1919Intro.htm

[gsgs]

Re: analysis by Frost, Sydenstricker

that took a while to digest.

that was all done without computers ...

they have data since 1862 for each camp, but only yearly.
Presumably there is a lot of travel between the camps, so I'm not
sure, how good the data would be to search for patterns
for specific virus strains.

[Niko]

Re: analysis by Frost, Sydenstricker

One of the charts from Jonesie

I FOUND ANOTHER ANALYSIS BY SELWYN D. COLLINS, Ph.D.
(US NAVY)

It has a LOT of charts, which I don't know how to post. Maybe someone could start another thread under Collins name?
It has nice charts for Influenza/pneumonia, 1887-1956:

http://www.history.navy.mil/library/...za_collins.htm

[gsgs]

Re: analysis by Frost, Sydenstricker

maybe an extra thread for charts only ?

here is another one:

http://www.pnas.org/content/vol102/issue31/images/large/zpq0310590700001.jpeg

(thanks Anne)

which seems to show that the 1915,1916 flu-outbreaks were not
related to the 1918 pandemic.

They were a bit unusual, though.

Apparantly there are no age-specific monthly P+I data for cities in that period.
Only age-specific all causes or yearly P+I.

Took them only 87 years to figure out that the 1918-pandemic was already in
February in New York. No computer needed.

[Jonesie]

Re: analysis by Frost, Sydenstricker

A CHART BY W.H.FROST:

http://www.ibiblio.org/hyperwar/AMH/...Pandemic-1.jpg

AND MORE CHARTS:

http://nmhm.washingtondc.museum/collections/archives/agalleries/1918flu/1918flu.html

[gsgs]

Re: analysis by Frost, Sydenstricker

yes, I found some more data, e.g. monthly from 1918,1919
or state - data.

But maybe OCR is possible and let me wait what the library people say,
what I can get...
They definately have deaths from all causes, maybe even weekly ,
that's useful for 1918. But for the other years I want influenza/pneumonia
deaths.




----------------------------------------------

http://statelibrary.dcr.state.nc.us/iss/INFLUENZA/PDFfiles/C340919.pdf#Table10

page 2 , cities - table --###done
page 3f, together with 28f : deaths per state rural vs. cities, influenza and all deaths , 4 numbers per state
page 12-13 , total deaths 1919 by month and city


same for 1918 :

http://statelibrary.dcr.state.nc.us/iss/INFLUENZA/PDFfiles/C340918.pdf#Table10

presumably other years are in the library too

http://statelibrary.dcr.state.nc.us/iss/INFLUENZA/PDFfiles/USPHR47All.pdf

page 6
6 cities, autumn 1918, weekly

http://statelibrary.dcr.state.nc.us/iss/INFLUENZA/PDFfiles/USPHR35All.pdf

page 4f , cases 1920 by city, weekly
page7f , excess influenza by cities, 1918, 12weeks



age and sex and races data not yet considered. Could be useful to determine whether the
virus has changed, how much pre-1918 was similar to 1918 or nowadays H3N2

http://statelibrary.dcr.state.nc.us/iss/INFLUENZA/PDFfiles/USPHR34JuneDec.pdf

page5 , weekly 1918, 13 British cities
page6 , weekly 1918 , 14 European and Indian cities
page13 , weekly 1917,1918 , London,Paris,Dublin


compare with weather data ! Or other weather dependent disease



influenza+pneumonia is preferred, not just influenza.
Some cities include influenza-->pneumonia deaths
in the influenza statistics, some in the pneumonia statistics.

Collins S, Frost W, Gover M, Sydenstricker E. 1930. Mortality from Influenza and Pneumonia in 50 Large Cities of the United States 1910?1929. Public Health Rep. 45:2277?2328.


sorry for the bumping ..., but I could no longer edit that post.
Added the last link, someone can access that ?
51 pages, 30 years, 50 cities,
monthly data would be 18000 numbers, typing one number per second
would take 5hours :-(

[gsgs]

Re: analysis by Frost, Sydenstricker

mortality in 44 US-cities per 1000 population in 2nd,3rd,"4"th wave:
(5th wave in 1920 was more severe then 3rd wave 1919 !)
69,01,03,Philadelphia
59,05,04
59,12,06
57,03,00
55,02,02
55,16,12
50,12,00
49,13,00
49,21,00
48,03,02
47,10,04
45,12,00
44,10,03
42,31,02
39,12,00
38,11,04
38,13,03
35,18,02
33,02,03
33,22,01
32,09,04
30,09,08
27,11,04
27,26,01
25,02,09
25,12,03
25,27,08
24,32,07
22,13,11
22,20,00
19,04,14
19,13,02
19,15,07
18,22,03
17,02,00
17,11,07
16,18,02
15,09,08
15,15,00
15,18,03
12,18,04
11,13,02
07,13,00
04,12,04,Toledo

this shows more deaths in later waves in cities who did well
in the 2nd wave. The relationship is D(3)+D(4)~-0.15*D(2)+22.
where D(i) is the deathrate per 1000 in wave i .

So 15% of spared deaths (compared with e.g. Philadelphia)
come back as real deaths (in average)
in later waves. Including 1920 and later may exceed
this to maybe 30% or more. And this ignores people who move.

It would be interesting to include 1920 here, also the 1st wave in 1918,
also the 1916 outbreak and to distinguish by ages.

[gsgs]

failure to infect "volunteers" in 1918

in spite of all evidence pointing to the importance of contact, attempts to transmit the disease experimentally under controlled conditions have uniformly failed. The United States Public Health Service sponsored two experiments of this character during the fall outbreak in 1918.31 One experiment was carried on in Boston, with 100 volunteers from the Navy, of the most susceptible age. None were known to have had influenza previously. These men were treated with influenza bacilli, with nasopharyngeal secretions, with and without filtration, by intranasal sprays, and by direct swabbing from patient to volunteer. The attempt was made to induce the disease by the injection of citrated blood from patients and the injection of filtrates of nasophyaryngeal secretion. Finally these men were exposed to the most intimate personal contact with patients in wards, all with the complete failure to produce the disease. A similar experiment was conducted with the same result in San Francisco. The explanation of this result is lacking. either the proper method of transmitting the disease was not used, which seems very unlikely in view of the diversity of methods employed, or the volunteers themselves were immune to the disease either naturally or through previous infection in spite of their negative history. This latter hypothesis is hardly satisfactory though seemingly more probable than the former.

31. Rosenau, M.J.; Keegan, W.J.; Goldberger, J.; and Lake, G.C.: Some Interesting Though Unsuccessful Attempts 6to Transmit Influenza experimentally, Public Health Reports, 1919, xxxiv, No. 2, 33. Also, McCoy, G.W., and richey, De Wayne: San Francisco Experiments. Public Health Reports, 1919, xxxiv, No. 2, 34.

http://statelibrary.dcr.state.nc.us/iss/influenza/PDFfiles/USPHR011019.pdf

SOME INTERESTING THOUGH UNSUCCESSFUL ATTEMPTSTO TRANSMIT INFLUENZA EXPERIMENTALLY.How great are the difficulties surrounding the study of the natureof the virus of influenza is indicated by the following summary of twoseries of experiments recently carried out, one at Boston and one atSan Francisco.'BOSTON EXPERIMENTS.These experiments were carried on jointly by Lieut. CommanderM. J. Rosenau, Medical Corps, U. S. N . R . F., and Lieut. W. J. Keegan,Medical Corps, U. S. N. R . F, and by Surg. Joseph Goldberger andAsst. Surg. G. C. Lake, United States Public'Health Service, at theUnited States Quarantine Station, Gallop's Island, Boston, Mass.The subjects of experiment were 68 volunteers from the United StatesNav,1 Detention Training Camp, Deer Island, Boston.These volun-teers had been exposed in some degree to an epidemic of influenza atthe training camp or at some station prior to coming to Deer Island;47 of the men were without history of an attack of influenza duringthe recent epidemic and 39 of these were without history of an attackof such illness at any time during their lives.The experiments consisted of inoculations with pure cultures ofPfeiffer's bacillus, with secretions from the upper respiratory passages,and with blood from typical cases of influenza.The study was begunon November 13 with an experiment in which a suspension of a-freshlyisolated culture of Pfeiffer's bacillus was instilled into the nose of eachof 3 nonimmunes an(l into 3 controls who had a history of an attackin the presentepidemic.None of these volunteers showed hnyreaction following this inoculation.Another experiment was madeat a later date with a suspension of a number of different pure culturesof Pfeiffer's bacillus, of which 4 were recently isolated.Ton pro-sitmably noninunune volunteers were inoculated, with the samene ative results.Three sets of experiments were made with secretions, both unfil-tered and filtered, from the upper respiratory tract of typical cases ofinfluenza in the active stage of the disease .In these experiments atotal of 30 men were subjected to inoculation b3. jucans of spray,075W-19-1(33)4

Page 2
January 10, 1919.34swab, or both, of the nose and throat . The interval elapsing betweensecuring secretions from the donors and inoculation of the volunteerswas progressively reduced in these experiments so that in the thirdof the series the interval at most was 30 seconds. In no instance wasan attack of influenza produced in any one of the subjects. Anexperiment was made in which the members of one of the groupsofvolunteers which had been subjected to inoculation with secretionswere exposed to a group of cases of influenza in the active stageof thedisease in a manner intended to' simulate conditions which in natureare supposed to favor the transmission of the disease. Each of thisgroup of 10 volunteers came into close association for a few minuteswith each of 10 selected cases of influenza in the wards of the ChelseaNaval Hospital. At the time the volunteers were exposed to thisinfection the cases were from 10 to 84 hours from the onset of theirillness and 4 of, them were not over 24 hours after the onset. Eachvolunteer conversed a few minutes with each of the selected patients,who were requested to, and coughed into the face of each volunteerin turn, so that each volunteer was exposed in this manner to all 10cases. The total exposure amounted to about tliree-quarters of anhour for each volunteer. None of these volunteers developed anysymptoms of influenza following this experiment.Advantage was .taken of the opportunity for making this study toattempt to confirm the reported positive results of transmission ofinfluenza by Nicolle. Secretions from 5 typical cases of influenzawere secured, filtered, aitd some of the filtrate was inoculated subcu-taneously into each of the group of 10 volunteers. At the same timeblood was drawn from_ the saii!.e cases and pooled, and some of themixed blood injected subcutaneously into each of another group of10 volunteers. The time lost between drawing the blood and inocu-lating it in no case exceeded three-quarters of an hour. None of themen subjected to these inoculations developed any evidences ofillness.In the foregoing experiments the patients serving as donors be-longed to groups from epidemic foci either on shipboard or, at insti-tutions . The great majority indeed belonged in a group from anepidemic on board the U. S. S. Yacona. Of the personnel of thisvessel, 95 in number, 80, or 84 per cent, were stricken with the.disease in an epidemic between November 17 and 29. ,SAN FRANCISCO EXPERIMENTS.The following observations were carried out practically simul-taneously with those described above. The work was done at theAngel Island Quarantine Station, San Francisco, Cal., utilizingvolunteers from the Ierba Buena Naval Training Station, SaarFrancisco. The experiments were carried on jointly by Surg. G. W.

Page 3
35January 10, 1 1)19 .McCoy, of the United States Public Health Service, and Lieut . DeWayne Richey, United States Navy. The volunteers who were usedin these experiments differed from those used at Boston in two re-spects-first, the personnel of the Yerba Buena Station had not beenexposed to influenza in the present epidemic and were therefore pre-sumed not to possess any special natural immunity ; second, all ofthe men had been vaccinated with large doses of a bacterial vaccinecontaining Pfeiffer's bacilli, the three fixed types of pneumococciand haemolytic streptococci. We are not prepared at present tostate what influence this vaccination may have had in promotingresistance to influenza infection, but if we may judgeT>y_the resultsof controlled experiments elsewhere such vaccination may for thepresent purpose be ignored.Brief details of the experiments are as follows :Work with Cult-arcs.A group of 10 volunteers was divided into 2 equal squads. Onegroup had instilled into the nostrils of each man a heavy suspensionmade by emulsifying cultures of S strains of Pfeiffer's bacillus withoutfiltration . The other group had the same material used after passagethrough a Berkefeld N candle . The results were negative, thoughthe men were held under observation for seven days.Work With Secretions.Four groups of volunteers, of 10 nien each, were used for theseexperiments. Emulsions of secretions from the upper respiratorypassages of active cases of influenza from 15 to 48 hours from theonset were instilled into the nose by means of a medicine dropper orwith an atomizer. In each experiment approximately an equal num-ber of volunteers were treated with the same emulsion after filtrationthrough a Berkefeld N candle. In every case the results were nega-tive, so far as the reproduction of influenza is concerned. The menwere all observed for 7 days after inoculation. In 3 cases in whichunfiltered material had been instilled sore throat developed whichcorresponded clinically with acute tonsillitis, and in two of thesecases an almest pure culture of a haemolytic streptococcus was securedfrom throat cultures.A filtered emulsion of material from the upper air passages of anacute case of influenza was dropped into the conjunctivae of 2 vol-unteers and the same material injected subcutaneously into _1 volun-teer. In each case the result was negative.One cubic centimeter of blood taken during the active stage of in-fluenza*cvas inoculated subcataneously into 1 volunteer with negativeresults.

Page 4
January 10, 1919.36In all of these experiments the time betweenthe collection of thematerial from the patient and its inoculation into thevolunteers wasin the neighborhood of3 or 4 hours. The conditions underwhch it -vas necessary to conduct experimentsdid not permit of ashorter interval . The unfilteredsuspensions which were used weresubmitted to cultural examination afterinoculation and found tocontain living organisms as follows: Pfeiffer's bacillus, piieumococciof group 4, and hemolytic streptococci.EDITORIALNOTE.-The foregoing experiments, though extremelyinteresting, do not, of course, warrant final conclusions. It is hopedthat it may be possibletocarry. the studies further and that resultsmay be obtained that will definitely settle thenature and the modeof spread of the virusofepidemic influenza. For the present thesanitarian will do well to continue to apply the general principles ofcontrol that are based on the justifiable assumption that the diseaseis a droplet infection, giving, however, increased attention to a pointthat is suggested by these experiments-namely, an ilnfective periodat the very earliest stages of the Attack.It would seem to be wise to give renewed emphasis to the impor-tance of going to bed at the slightest indications of illness.

[gsgs]

Re: 1918 Analysis with Charts and Graphs

here is a 1000 pages free online book with statistical tables from 1900-1940.

80 MB, .pdf file, so think twice whether you click on the following link !

http://www.cdc.gov/nchs/data/vsus/vsrates1900_40.pdf

someone wants to type tables into computer ... ?
It should be possible to get the data with OCR,
since Adobe-reader allows searching it for keywords.
It's not in the online google-cache, but google can search it
for keywords/keynumbers.
Can someone convert pdf to plain text ?


the picture above with the 1918 CFR in US-cities, I could try
to write a script to put it into Hawkeye's map, if there is interest ?


----------------------------------------
page 254 , table 14 ,

deaths per 1M population in USA of influenza and pneumonia
age <1,1-4,5-14,15-24,25-34,35-44,45-54,55-64,65-74,75-84,>84
years 1900,..,1940

  2022 16821  3866   382   481   762  1141  1752  3294  6924 15313 28995
  1972 15203  3153   344   493   769  1122  1689  3327  7246 17468 36102
  1613 16508  3563   315   374   564   841  1272  2322  4775 10490 19161
  1693 15400  3088   306   378   616   944  1432  2690  5673 13279 25510
  1921 16457  3320   321   423   722  1174  1713  3221  6560 14717 32336
  1693 15986  2964   275   350   570   946  1420  2719  5830 13498 29226
  1563 16499  3249   264   332   514   829  1260  2395  4925 10892 21680
  1800 16207  3103   287   370   627  1007  1527  2975  6635 15700 32789
  1509 15186  2759   264   294   450   755  1247  2380  5360 13036 26139
  1481 16198  2891   246   268   427   736  1130  2288  5093 11668 23601
  1559 16290  2856   254   292   469   788  1257  2502  5566 12928 25550
  1454 14093  2442   233   286   430   739  1206  2354  5517 13251 27922
  1384 14910  2362   202   261   416   694  1108  2199  4933 11724 23965
  1408 15025  2503   232   272   424   732  1129  2148  4837 11602 24150
  1324 14398  2148   219   255   387   675  1102  2115  4601 10963 22446
  1459 14894  2055   218   280   430   710  1211  2400  5679 13715 29304
  1633 14140  2217   230   350   514   877  1404  2773  6613 16315 36988
  1645 14745  2115   240   389   593   981  1488  2814  6146 15030 31874
  5885 22733  7160  1762  5805  9926  5548  3478  3819  6463 11790 22306
  2230 15942  2939   633  1414  2359  1810  1639  2332  4596  9139 18422
  2073 14950  2837   452  1013  1806  1640  1649  2558  5459 11943 23791
   987  9795  1468   220   263   327   516   782  1455  3354  8426 18419
  1323 12650  1900   240   370   512   777  1081  1934  4571 11260 25303
  1517 13566  2007   260   435   560   879  1269  2287  5498 14319 34740
  1152 11848  1518   200   322   410   679  1031  1704  3799  9530 23613
  1217 11751  1503   233   361   443   741  1043  1778  4049 10716 27585
  1417 14005  1961   230   354   504   839  1227  2045  4700 12969 32454
  1022 10160  1291   188   273   374   624   922  1530  3400  8736 21688
  1425 13215  1766   275   428   566   916  1284  2037  4693 13055 33039
  1465 13219  1791   258   410   583   926  1322  2177  4967 14071 34923
  1025 10901  1231   188   278   389   611   957  1646  3370  8343 18844
  1075 11054  1253   196   304   446   678  1010  1649  3404  8840 21747
  1073 10305  1158   204   305   425   652   955  1629  3484 10065 26162
   957  9900  1090   175   250   380   597   884  1493  3005  8362 21688
   969 10553  1154   178   246   363   576   923  1571  3025  8029 19002
  1042 10029  1116   199   286   436   666  1016  1669  3259  8535 20718
  1196 10904  1193   218   356   501   778  1189  1941  3722  9776 23302
  1149 10364  1124   179   294   458   721  1169  1898  3576  9603 24259
   804  9019   902   137   184   270   423   709  1233  2482  6775 15921
   757  7937   703   114   152   222   371   635  1137  2456  7170 19301
   703  8701   625    90   115   171   236   518   937  2264  6912 19299

[wetDirt]

Re: failure to infect &quot;volunteers&quot; in 1918

...much snipped...

For the present thesanitarian will do well to continue to apply the general principles ofcontrol that are based on the justifiable assumption that the diseaseis a droplet infection, giving, however, increased attention to a pointthat is suggested by these experiments-namely, an ilnfective periodat the very earliest stages of the Attack.It would seem to be wise to give renewed emphasis to the impor-tance of going to bed at the slightest indications of illness.

And THIS is why the rate of infection of healthcare workers sas/is so low. The infectious window had already passed, when the patient might not have even known they were infected, or infectious for that matter. As I have said repeatedly, and nobody seems to pick up on, HCW are in more danger talking to family members of an infected person, than from the sick person themself. That's why staying home is so important. You can't **see** the infectious people. Even the use of heat scanners might be insufficient, because it's not clear that any study has linked the infectious period to an actual rise in body temperature. It's still all just **assumed** that there is a rise in temperature that coincides with the infectious period.

[gsgs]

Re: 1918 Analysis with Charts and Graphs

infectious period for influenza is given as 4.1 days,
average latent period 2 days. First symptoms appear in average
after 2 days. So typically(average) you would be infectious from
onset of symptoms and then for 4 days.

[gsgs]

Re: 1918 Analysis with Charts and Graphs

US-death rates by month 1900-1940, total, Jan,..,Dec.




you can see that deathrates in May,June,..,Nov. are pretty stable except 1918.
The seasonal changes in Dec,Jan,Feb,Mar,Apr are most due to influenza.
One Polio(?)-outbreak in summer 1911.

Also, the virus before 1918 caused a bit less predictable changes
and the deathrate went down until 1918, from 1918-1940 it decreased more slowly.

[gsgs]

Re: 1918 Analysis with Charts and Graphs

I found mortality data from Sweden since 1751 !

There was crop failure in 1771 and 1772 and famine 1772,1773
which was 4 times as severe in Sweden as Spanish flu in 1918 !

Finnish war and epidemics in 1808,1809.
Data before 1860 is less reliable, see:

These events were much worse for Sweden than Spanish influenza
in 1918 with excess deaths of 0.66&#37; (USA:0.39%).

Excess mortality also in 1892f and 1899f , this could be due
to influenza pandemic 1890-1893 and 1899-1901, I'm not sure.

>It reached Sweden in late November 1889 and swept through the country
>during the following months, mainly following the major transport routes,
>especially the railways. The morbidity was described to be between 50 and 60%
>with an increase in mortality during the actual pandemic weeks of 250% of the average.

however total mortality in Sweden in 1889,1890 was low.

I couldn't find the reason for the excess deaths in Sweden in 1857.

1783 was the eruption of Laki-volcano which resulted in bad
crops in Europe the following years.





(children below 5 years are excluded because of high mortality in 18th and 19th century)


data from:
Human Mortality Database. University of California, Berkeley (USA), and Max Planck Institute for Demographic Research (Germany). Available at www.mortality.org or www.humanmortality.de (data downloaded on [date]).
the original Swedish sources are listed here:

http://magictour.free.fr/sw1.pdf