Using German Medical Data to Track Violence on the Western Front

[Sasho Todorov]

Part I: Overview

 

Thanks to the kind help of Great War Forum members Phil Andrade, and the original labor of member Paul Hederer, I've been able to get my mitts on a scanned copy of the Sanitatsbericht uber das Deutsche Heere, the German official medical study of the war. It's a treasure trove of information, including much that the famous Blood Test Revisited ignored (such as the number of German soldiers who died of their wounds). Among these are the monthly army reports of woundings by cause, as well as sicknesses. For some reason it does not contain the KIA and MIA reports, because apparently it was more useful to have a diptheria table than a KIA/MIA table. Only mildly bitter about that, especially thanks to the way it teases the reader by providing the 10 day numbers from certain battles.

 

Anyway, the Sanitat breaks down its wound reports into four categories: gunfire (including artillery), armes blanches, broken bones+sprain+dislocations, and other wounds. As mentioned above, it does so by army and by month, giving a fairly granular access to the data. In addition, it also gives the ration strength of each army for each month, instead of its paper strength. This is naturally extremely useful for not only tracking total numbers, but also what percentage of a force was lost in such a manner. Unfortunately the Sanitat only officially measures its monthly reports through July of 1918, though it actually covers the remaining period of the war in its overall data analysis (including, which the Blood Test did not notice, a completely KIA+MIA presumed dead total of 2,037,000). 

 

I began this project by examining whether or not the four categories should be further divided. In essence, only gunfire and armes blanches wounds can be directly linked to hostile Entente action, the rest are potentially linked to accidents. This factor can be seen when you compare the two charts of total wounded (above) and what I call direct violence wounds (below). Keep in mind that this is not an MIA+KIA+Violence Woundings total, it's only the woundings here, with the goal being to primarily analyze trends. 

 

 

 

While my current chart formats unfortunately obscure it somewhat, you can tell here that there's a more substantial "floor" to the total woundings than there is to the direct violence wounds. My belief is that this represents the floor of wounds emerging due to accident or other causes. This is backed by statistical analysis of the direct (gunfire and blades) and indirect (broken bones and other wounds) categories. 

 

After doing a correlation coefficient calculation, the R between the number of Total Western front direct monthly woundings and indirect wounds was only .418. The data points measured were total monthly direct and total monthly indirect losses. Army R coefficients were also measured, though I'm still working on the data there. To give some context, R measures correlation, a perfect 1 means for every increase in one category you see an equal in another. To give a sense of scale, below an R of .3 is considered a weak correlation, from .3-.7 is moderate, and .7 to 1 is strong. What this means is that the correlation between the direct hostile woundings, and the mixed category of accidental and hostile caused broken bone and other woundings, categories is only moderately weak. This applies in both ways as well, as it could show periods of significant increases in non-direct violence wounds as compared to violence wounds, or vice versa.

 

Therefore, I will primarily be using my direct violence chart for further analysis in this post. For ease of interpretation, I have actually organized each section, top to bottom, by its location from East to West on the Front. This is why, for example, the Orange section, representing the 6th AOK, flips from 2nd to the Bottom to 2nd from the Top in October of 1916, as it depicts the Army's transfer to from Alsace-Lorraine to the Arras sector. Of particular note is the way in which the 1st Army (depicted in red) shifts, as it is deactivated in October of 1915, before being reactivated for the Somme in July of 1916, and then shifted to the Champagne sector of the Nivelle Offensive in 1917. 

 

I will also be using the chart below, which is a percentile based analysis, rather than a total number analysis, for direct violence woundings on the front. Basically, it measures the portion of each months total direct violence woundings that a particular army's sector is responsible for.

 

 

Edited 31 May , 2020 by Sasho Todorov

[Sasho Todorov]

Part II: Takeaways of note.

 

For the purpose of this section, I advise having the charts open, if possible, to follow along.

 

The first, and most obvious, takeaway is that the Western front follows a clear "calm season" between major periods of campaign activity. This is except for 1915, where the fighting does slow down, but maintains a far more active pace during the winter than in subsequent years. This represents the often forgotten period of Joffre's first winter offensive, especially along the Champagne-St. Mihiel portion of the front. The calm sector depends in length, with three months from 1915-1916, broken by the Germans at Verdun, four months in 1916-1917, broken by the launching of the Entente spring offensives of 1917, and then only two months from January to February of 1918 before the Spring Offensives.

 

Secondly, the scale of the Spring Offensives are truly brought to light here, with wound rates 60% greater than even July of 1916, when the Somme and Verdun were simultaneously raging.

 

Third, and this is particularly important given the classic narrative of the war, the French army does not go dormant following the mutinies. While British responsibility for the war effort clearly takes some primacy, the French sector (roughly measured by the line dividing the 2nd--green--AOK and the 7th--blue--AOK) remained responsible for German direct violence wounds at a percentile rate of between the mid-40's and very high 30's. While clearly lower than before, this is also because of the much greater burden assumed by the French during the war. Even in the initial offensives of 1917, Arras (in orange) is responsible for a significantly smaller percentage of German monthly losses than the Nivelle Offensive. A high rate of violence continues in the Chemin des Dames as well, with the oft-forgotten period of German offensive action, the battle of the Observatories, in June and July of 1917 accounting for nearly as many violence woundings as the individual months of April or May. Clearly there was a slow down, but this is hardly a "wastage" period.

 

Fourth, continuing the trend of 1917, the 2nd Battle of Verdun is much more significant than usually presented. In August of 1917 the German 5th army suffers about 50% of the wounding losses as the 4th army at Ypres, and in September the fighting continued at a strong pace, with losses about 40% those of Ypres. The peak of British predominance in direct violence woundings in 1917 doesn't hit until October of 1917. This is one of the rare months where the lack of MIA and KIA data for each army greatly complicates analyses, because of the nature of the battles of La Malmaison, which saw an extremely high proportion of German KIA and MIA due to the collapse of the forces there. The classic problem here is comparing it to Broodseinde, which also saw a high KIA and MIA rate, but did not see the equivalent collapse in the German lines, as can be ascertained by the much lower BEF casualty trade coefficient.

 

Fifth, Champagne east of the hills is a remarkably quiet sector for much of the war between the failed attempts to break through there in September and October of 1915 and its reheating during the final German offensive of 1918.

 

Part III: Conclusions:

 

I hope these charts have been of some use for visualizing rates of violence during the war. As noted, the lack of MIA and KIA data greatly complicates matters. However, I do believe this tracking of gunshot and artillery woundings gives a fairly good impression of where violence was occurring amidst the front. I am currently in the process of OCR'ing the Sanitatsberich, as well as transcribing its tables, and I'll have a larger announcement on that project, including access to its data, ideally by July.

Edited 31 May , 2020 by Sasho Todorov

[phil andrade]

Sasho,

 

 

This is groundbreaking stuff .

 

There are some figures for wounded in the tables of the sanitatsbericht that are so huge - compared with relatively tiny numbers of killed for the same period- that  I’m convinced that they conflate accidental non battle injury with battle wounds at a time of low activity.

 

The original Blood Test chapter, written by Winston Churchill in the decade after the Great War, is a must read for a comparison with the graphics that you’ve so painstakingly compiled.

 

The performance of the French in the second half of 1917 is something that demands to be reconsidered , and your research is invaluable here.

 

 

Phil

 

 

[phil andrade]

Sasho,

 

As an afterthought to my previous post, I provide data on German losses in killed and missing in action that might be useful as collaborative evidence for the conclusions we might draw from your graphs.  These figures are from Appendix X of Churchill’s History of The Great War.  These are Reichsarchiv tabulations of Losses on the Western Front by Main Operation Periods.

 

The figures are for battle casualties in  killed and missing only, and do not include deaths from wounds or other causes. The missing include prisoners.  There were roughly a million Germans wounded in action in addition to these.

 

For the period July to December 1916, there were 145,295 suffered on the French front, and 81,968 against the British.  The French lost 144,000 and the British 125,575.

 

What a   remarkable disparity in the respective exchange rates.

 

Between January and July 1917, the French accounted for 93,474 German killed or missing, the British 82,552. French and British, 107,000 and 95,742 respectively.

 

From August to December 1917, the figure on the French front was 59,276 ; against the British it was 89,478. Against these, the French figure was 38,000, and the British 92,459.  The disparity in favour of the French is dramatic here.  Malmaison comes to mind.  The British were getting into their stride, too.  Think of Currie's success at Lens.

 

Phil

Edited 31 May , 2020 by phil andrade

[Sasho Todorov]

The Blood Test data from Churchill is indeed fascinating, I think it gives a great starting point for general proportions and ratios, and one of my main goals with my work on the Sanitat, especially in the r coefficient and other deeper statistical work, is to try and get to the point of knowing what we can and cannot verify from it. Having that bloody MIA and KIA army monthly data would solve the issue right then and there, but for some Godforsaken reason the Germans thought it wasn't worth including (but scarlet fever, by contrast, with its all of 2,000 total cases, of course got its own table. There's a bureaucrat in the afterlife that I'm going to have a very stern talking to with once my time has come.)

 

The Sanitatsbericht is fascinating in this regard in that it actually addresses the very issue touched on by Churchill in his footnotes, the problem of the Reichsarchiv data as a "entry point" measure (i..e. the idea it doesn't tell you how many people died of their wounds). What makes the Sanitat so fascinating is that it not only provides the entry point data, but also has more general data as to patient outcome. As mentioned before it explicitly states a death toll for the entire war of 2,037,000, and it's a substantially more modern tabulation than the ZentralNachweisant data. The final data provided on page 12 is: 1,900,876 dead, 4,215,662 WIA, and 974,977 POW+MIA not confirmed to have died. It even suggests that 100,000 of the remaining POW+MIA count needed to be transferred to the dead, which is where the 2,037,897 dead citation originally stems from (the balance being Kriegsmarine and colonial deaths).  It's why, in retrospect, I'm extremely frustrated by the authors of The Blood Test Revisited for just restating a couple of tables and not even skimming the work, as they give outdated Zentral data. 

 

I share your suspicions about accidental woundings being included in total woundings, it's why the low R coefficient between the categories of wounds that must be the product of Entente action, and those that may be otherwise, is so important. The floor of woundings for even quiet sectors without direct violence woundings simply can't occur without at that being the results of accidents. Unfortunately, I am having some difficulties getting through the signal noise with a study as a whole. My current pet project on this is to segment out calculations of R into time period blocks defined by having a sufficient level of violence as to qualify as major fighting. Luckily only 1/3 of woundings fall into question, and, even then, the bulk of these are going to be linked to direct hostile action. 

 

To address another problem of the very high wound rate compared to others, the Sanitat also has very detailed tables on Section II, page 128 (tables 124-127) which cover the annual data for inflow and outflow for each cause from military hospitals, including the reasons why the patient left the hospital (returned to the colors, died of wounds, otherwise). This is further tabulated in table 151, at the end, which gives a very detailed account of manpower loss on both fronts, though it unfortunately treats the hospital system as a single entity unattached to either Front. I'm still working on tabulating and going through the details there, but Churchill's addition of 300,000 died of wounds is not entirely off base (the real total is 289,053 for both fronts, of which 75% likely is from the Western Front).

 

One thing I have been able to work on is this, though (pre suggested final transfer of 100,000 MIA from the German POW+MIA to the Dead column). Using the ratios of percentage of losses by cause provided within the Statistics of the Military effort, I was able to work out the specific BEF total figures. French figures from Churchill (and likely from the Official French History).

 

	Died	Wounded	POW+MIA	Total
Germany	1900876	4215662	974977	7091515
France	1356000	3178420	513080	5047500
UK	725000	1843000	172000	2740000

 

More interestingly is the proportions:

 

Proportions
Germany	26.80%	59.45%	13.75%
France	26.86%	62.97%	10.17%
UK	26.46%	67.26%	6.28%

 

Both Germany and France having almost identical ratios of dead, and the UK being barely half a % behind, should I think lay rest to the classic argument about whether or not the Germans counted the lightly wounded in their Sanitat figures. The core difference is mostly in the ratio of POW to Wounded. This is likely down to the proportion of time each party was on the offensive (with the French being particularly badly hit by the major POW loss at Mauberge in 1914), with the differences coming down to whether one party's wounded was behind one's lines or in enemy territory when it came time to bring them back from the battlefield or not.

 

 

Edited 31 May , 2020 by Sasho Todorov

[phil andrade]

Sasho ,

 

The sanitatsbericht is, I suppose, a didactic work, with a plethora of details about sickness and accidental injury which, it must be said, accounted for a huge number of hospital admissions.

 

Imagine the number of soldiers who suffered accidents shifting sand bags and barbed wire etc during construction of those elaborate defences.  Yes, if you need to find out how many Germans suffered from piles during the Riga operations of September 1917, this is the place to come to : not so good on details of killed and missing, who, I suppose, are outside the remit of hospitals anyway.

 

I think the 289,053 died of wounds you allude to exclude all who died after the end of July 1918.  I wonder if it includes wounded German prisoners who died in Allied hands.

 

Churchill’s rendition of German losses on the Western Front was compiled with the help of a German official at the Reichsarciv in the 1920s. No excuses about the loss of records after bombing in 1945 !

 

While his figures are something in the nature of “ a wing and a prayer “, I reckon they’re pretty sound .  He reckons that German dead on the Western Front were just shy of one and a half million, and that their overall casualties there were c.5.4 million.

 

Evidence from the German War Graves lend credence here : they cite a figure of 132,000 German soldiers from 1914-18 buried in Belgium alone, with the proviso that another eighty to ninety thousand have not been interred and remain missing there.  If we apply that  proportion  to the overall number of 900,000+ interred in France and Flanders, then Churchill’s guess passes the sniff test.

 

Phil

 

 

 

 

 

[AOK4]

7 minutes ago, phil andrade said:

Evidence from the German War Graves lend credence here : they cite a figure of 132,000 German soldiers from 1914-18 buried in Belgium alone, with the proviso that another eighty to ninety thousand have not been interred and remain missing there.  If we apply that  proportion  to the overall number of 900,000+ interred in France and Flanders, then Churchill’s guess passes the sniff test.

 

 

 

The number of missing in Belgium is highly controversial in my opinion (and I would like to know how they come to this number). The Battles of Ypres were not only on Belgian soil from a German point of view: at least 10,000 (a very conservative estimate, probably more like 15 to 20,000) German soldiers either KIA or DoW lay buried in France. And I would add several hundreds of fallen (more probably a few thousand) repatriated to Germany, mainly during the war.

 

BTW, on 17 December 1957 there were officially 134,481 German burials in Belgium (excluding the 542 in Lommel and the odd 10 or so buried in private graves in Belgian communal cemeteries).

 

Jan

 

[phil andrade]

Jan,

 

Your eagle eye has caught me out !

 

The figure of German dead that I used was on display in a museum in Ypres : In Flanders Fields, if I remember correctly, is the name of the establishment.  I made the assumption that such a citation would only be made if it had been backed up by some authoritative source.  Indeed, I think that I have seen the figure of 210,000 to 220,000 mentioned elsewhere, maybe in some pamphlet produced by the German War Graves organisation.  It does seem quite plausible , given the difficulties faced by the Germans in recovering and interring their dead : if one quarter of all British dead on the Western Front were never recovered, then I would have expected that a higher proportion of German dead remain lost in all the battlefields of the Western Front, perhaps more so in Belgium than in France.  Wasn't it quite recently that eighty four German soldiers were discovered in the Wytschaete sector of the Ypres battlefield ?  Such large numbers do suggest that , to a far greater extent than their British counterparts, the German dead remain unrecovered. A diffident suggestion on my part...a bit of " a wing and a prayer", I confess !

 

Phil

[AOK4]

16 minutes ago, phil andrade said:

Jan,

 

Your eagle eye has caught me out !

 

The figure of German dead that I used was on display in a museum in Ypres : In Flanders Fields, if I remember correctly, is the name of the establishment.  I made the assumption that such a citation would only be made if it had been backed up by some authoritative source.  Indeed, I think that I have seen the figure of 210,000 to 220,000 mentioned elsewhere, maybe in some pamphlet produced by the German War Graves organisation.  It does seem quite plausible , given the difficulties faced by the Germans in recovering and interring their dead : if one quarter of all British dead on the Western Front were never recovered, then I would have expected that a higher proportion of German dead remain lost in all the battlefields of the Western Front, perhaps more so in Belgium than in France.  Wasn't it quite recently that eighty four German soldiers were discovered in the Wytschaete sector of the Ypres battlefield ?  Such large numbers do suggest that , to a far greater extent than their British counterparts, the German dead remain unrecovered. A diffident suggestion on my part...a bit of " a wing and a prayer", I confess !

 

Phil

 

Phil,

 

I'm writing some books about German cemeteries. I'm of the opinion that the Germans did bury most of their fallen unlike the British (the exception being the 1917 fighting and the retreat 1918). They held the fought over territory with most of the casualties from 1914/15 from mid 1915 to mid 1917 and did recover most of the fallen during this period. The matter of Wytschaete has to do with local circumstances in 1914/15 (bad roads so that evacuating was very difficult and Bavarian units who were happy to confirm the death of the person and his identity and then just burying them on the spot). The finding of the 80 something bodies there confirmed my gut feeling and this was further confirmed by some archival research.

 

One has to remember that the current German war graves service is only responsible for the First World War graves since after WW2 and that most of the archives (the Zentralnachweiseamt) were destroyed in 1944/45. Their knowledge of the evolution of cemeteries and graves is limited.

I know the number that they usually give, but my numbers come from their own archives...

BTW: the number is without any recent finds and reburials in the Kameradengrab (the exact number of reburials there seems unknown to the VDK itself).

 

Jan

 

PS: In the summer of 1918, hundreds of bodies were recovered from the recaptured Ypres battlefield. I hope to discover some sources about this, but for now, there's only a few snippets here and there in regimental histories.

Edited 31 May , 2020 by AOK4

[phil andrade]

Jan,

 

It’s all too apparent from the story of the Fromelles battlefield that the Germans went to greater lengths than their entente foes to recover and inter the dead.

 

This applied to fallen enemies as much as it did to their own men....the Pheasant Wood burials attest.

 

It’s been suggested that the Germans, as the invaders and occupying forces, had a vested interest in tidying up the battlefield, since it was their intention to hold and consolidate what they possessed, while the Allies might have regarded such activity as acquiescence in the status quo : an intolerable thing while Germans were holding so much Franco Belgian territory.

 

i suspect, though, that German discipline and hygiene practice had a lot to do with it.

 

That notwithstanding, isn’t it the case that a lot of German military cemeteries were destroyed in the war’s aftermath, with consequential loss of bodies ? 

 

The huge concentration cemeteries only allowed for a portion of recovery  ?

 

If two million German soldiers died in the war, would you endorse the suggestion that three quarters of them were casualties of the Western Front ?

 

Phil

 

 

 

 

[phil andrade]

On 31/05/2020 at 10:23, Sasho Todorov said:

 

 

One thing I have been able to work on is this, though (pre suggested final transfer of 100,000 MIA from the German POW+MIA to the Dead column). Using the ratios of percentage of losses by cause provided within the Statistics of the Military effort, I was able to work out the specific BEF total figures. French figures from Churchill (and likely from the Official French History).

 

	Died	Wounded	POW+MIA	Total
Germany	1900876	4215662	974977	7091515
France	1356000	3178420	513080	5047500
UK	725000	1843000	172000	2740000

 

More interestingly is the proportions:

 

Proportions
Germany	26.80%	59.45%	13.75%
France	26.86%	62.97%	10.17%
UK	26.46%	67.26%	6.28%

 

Both Germany and France having almost identical ratios of dead, and the UK being barely half a % behind, should I think lay rest to the classic argument about whether or not the Germans counted the lightly wounded in their Sanitat figures. The core difference is mostly in the ratio of POW to Wounded. This is likely down to the proportion of time each party was on the offensive (with the French being particularly badly hit by the major POW loss at Mauberge in 1914), with the differences coming down to whether one party's wounded was behind one's lines or in enemy territory when it came time to bring them back from the battlefield or not.

 

 

 

Sasho,

 

The proportions allow for differing interpretations.  It's my opinion that the higher proportion of British wounded to dead is attributable to the preponderance of their dead being casualties of the second half of the war, when the fatality rate among the men who were hit tended to diminish : look at the figures for 1918, and compare them with those of 1914, and you'll see what I mean.  The British figure for wounded includes gas cases, which represented quite a significant portion of the 1918 casualties, and resulted in a small ratio of deaths.

 

The French, on the other hand, suffered the greater part of their deaths in the earlier part of the war, and in this phase - especially in 1914- the ratio of killed among the casualties was high.

The figure of 3,178,420 wounded that  you cite is, I believe, inflated by several hundred thousand, because it contains all wound cases treated : multiple woundings, Allied and even enemy soldiers treated, and accidental injuries. The German numbers of wounded allows for the number of German soldiers  who were wounded and survived, as opposed to the number of wound cases treated.  I state this with some diffidence, and, I hope , constructively.

 

The soviet demographer Urlanis made some striking observations when he assessed the sanitatsbericht and applied his extrapolations to his survey of deaths in the Great War.

 

Contrasting the figures of 772,687 confirmed killed in action tabulated in the sanitats. with the overall total of 2,036,897 German military deaths that were finally acknowledged, he assessed the implications and concluded that 1,473,000 German soldiers had been killed on the field, 320,000 had died from wounds, 3,000 had died from gas poisoning..... and that disease, accident, suicides and the ordeal of captivity as prisoners of war accounted for the balance of 241,000.  Interestingly, he estimated that, of the 1,473,000 killed outright in battle, 1,100,000 were attributable to the Western Front.  This implies that 240,000 were mortally wounded there. All in all, a rough and ready endorsement of what Churchill reckoned.

 

When it comes to the figures of German PoWs on the Western Front, it's a remarkable feature that almost exactly half of all prisoners claimed by the Allies throughout the war were ascribed to the period from mid July 1918 to the Armistice.

 

ADDING AN EDIT HERE : From AFGG. Volume 3, page 602, French casualties, Western Front only, up until the end of 1915 :

 

Confirmed dead : 580,768

Missing : 406,026

Wounded : 945,257

 

Total : 1,932,051

 

Note that the confirmed dead alone equate to more than thirty per cent of the total casualties.  The actual deaths were understated, because many of the missing were obviously dead. A more accurate total for French fatalities on the Western Front, up until 31 December 1915, allowing for the missing who had been killed, is 643,322. ( Gray and Argyle, Chronicle of the First World War, Volume one, pages 282 & 285).  In this case, I suspect the total for wounded is understated...but my point is clear : in the earlier part of the war, the French casualties included a higher proportion of deaths than the later part.

 

Another edit : Contemporary official compilations show a good " snapshot" for the way that French and German casualties were comprised as of 31 December 1915 :

 

French Armies, all fronts:

 

Dead....590,407 ( 30%)

Wounded...960,416 ( 49%)

Missing....410,864 (21%)

 

German Armies, all fronts :

 

Dead...628,445 (24.7%)

Wounded..1,595,406 ( 62.7%)

Missing....320,154 (12.6%)

 

Even allowing for the large number of German missing who were dead, it's very clear that the French casualties were of a more fatal nature than their German counterparts.  The Battles of the Frontiers, August-September 1914, were the principal reason for this.

 

 

 

Phil

Edited 1 June , 2020 by phil andrade