Military History | How To Make War | Wars Around the World Rules of Use How to Behave on an Internet Forum
Iraq Discussion Board
   Return to Topic Page
Subject: Lancet study of Iraq war death toll - 655,000 deaths
eu4ea    10/11/2006 2:36:42 PM
The Lancet, one of the oldest and most respected peer-reviewed medical journals in the world has just published a report on the casualties produced by the war in Iraq to date. It was authored by the Jhons Hopkins School of Medicine, the School of Medicine of the University of Bagdhad and the Center for International Studies at MIT. Some highlights include: - The total death toll to date is 655,000 people - That comes to an average of 500 people/day - The death toll has has been increasing steadily since march 2003 - The data reflect only deaths above the normal death rate Here are links to CCNs story about this, the report itself (pdf format), the Lancet's site, and entries regarding the Lancet in wikipedia: / http://i.a.cnn.net/cnn/2006/images/10/11/human.cost.of.war.pdf / http://en.wikipedia.org/wiki/The_Lancet http://en.wikipedia.org/wiki/Medical_journal This comes on the heels of the July report by UN Assistance Mission for Iraq, which reports that an average of over >100 iraqui civilians were being killed every day. See The UN figures were obtained from the records from the Bagdhad morgue, and hence excluded large segements of the population (those killed outside of Baghdad, those who were not brought to the morgue after being killed, those kidnaped then killed, those killed then buried by their families, those buried in rubble, etc.) The reaction from the Administration, predictably, has been one of flat denial. Bush himself opined that "The methodology is pretty well discredited", despite the fact that: 1- Randomized large-scale household sampling is the 'gold standard' methodology used by modern epidemiology to track anything from flu epidemics to the results of natural disasters and armed conflict. The Lancet's implementation of this methodology was absolutely impecable. 2- The results are fully coherent with comparable conflicts (Vietnam, Congo, East Timor...) in terms of both the development pattern and the total death toll, roughly 2.5% of the population. 3- The Lancet has been around since the 1820's and is arguably *the* most respected medical Journal in the world (along with the New England Journal of Medicine). 4- This comes on the heels report by the UN Assistance Mission for Iraq on July 18th. That report, based on Baghdad morgue records, found that over 100 civilians are killed daily in Baghdad alone - and that's only counting those brought to the morgue as opposed to just being buried by their families. The only alternative Bushie had to offer was a 30,000 figure he pulled out of his hat last December, and is absolutely laughable. For your amusement, here's the clip of George Jr he-hawing and giving research methodology pointers to The Lancet: javascript:cnnVideo('play','/video/politics/2006/10/11/sot.bush.iraq.death.toll.study.cnn','2006/10/18'); When the hell are we going to start getting even a minimal degree of realism from this administration? You can agree or disagree with their policies - that's democracy and it's just fine. I would have much less of a problem with it if they came out and said "this is the reality but we still feel the invasion was the right thing to do" However, this flat-out denial of reality is simply not respectable. It's 'steadfast', it's 'not staying the course', its not inspiring and it's not respectable. Heart, eu4ea
 
Quote    Reply

Show Only Poster Name and Title     Newest to Oldest
Pages: 1 2
EW3       10/11/2006 2:55:53 PM
 
This belongs in the "October Surprise" thread....
If you go to google you'll find a similar story showed up on 29 Oct 2004 on CNN.  http://www.cnn.com/2004/WORLD/meast/10/29/iraq.deaths/ ; I guess this is why these are called "October Suprises". 
It too was done by lancet, that so reliable source, using the same statistically questionable technique.  Last time it was only 100,000 dead. This year it is 640,000.  Have to ask how mre than a half million people go missing in 2 years in a country with media people running all over the place looking for things to blame the US for.
Only truely mentally impaired left wing moonbats buy this story, and that is because they want to believe anything bad about the US. 
This is the kind of story that makes me which libel laws weren't so loose. 
 
 
Quote    Reply

eu4ea       10/11/2006 4:15:30 PM

This belongs in the "October Surprise" thread....

If you go to google you'll find a similar story showed up on 29 Oct 2004 on CNN. http://www.cnn.com/2004/world/meast/10/29/iraq.deaths/ ;" target="_blank">link I guess this is why these are called "October Suprises". 

It too was done by lancet, that so reliable source, using the same statistically questionable technique.  Last time it was only 100,000 dead. This year it is 640,000.  Have to ask how mre than a half million people go missing in 2 years in a country with media people running all over the place looking for things to blame the US for.

Only truely mentally impaired left wing moonbats buy this story, and that is because they want to believe anything bad about the US. 

This is the kind of story that makes me which libel laws weren't so loose.

EW3?  Not something I expected from you.  Maybe there's something I'm missing - either regarding the report itself or regarding the amount of vehemence around it.

Before we get much further, did you actually read either the Lancet's report or it's appendix on methodolgy and charts? 
They are here:  i.a.cnn.net/cnn/2006/images/10/11/human.cost.of.war.pdf
and here:  i.a.cnn.net/cnn/2006/images/10/11/appendices.human.cost.of.war.pdf

Both of them are .pdf files, and you'll have to copy &paste that to your browser - I still cant get this new system to present links properly.  If you havent yet, I do recommend you read them; the Lancet doesnt publish crap - and this study's methodology, implementation and reporting is particularly impecable by any standards.

As for your points -

1- Re: the credibility of the Lancet.  Well, simply put, the Lancet is one of the most respected peer-reviewed scientific journals in the planet, and has been so for almost 200 years now. This is no newspaper or hack and they dont publish trash - you'd have to look at things like the New England Journal of Medicine or the Proceedings of the National Academy of Sciences to find anything of comparable credibility.

2- Re: the 2004 report.  Yes, absolutely.  Not only do they acknowledge it, it's explicitly referenced in the introduction. As far as I can tell, finding 100k excess deaths in year 1 of the war is fully consistent with finding 655k in years 1, 2, and 3 combined. In fact that closely mirrors other reports of steadily increasing mortality in the Iraq war. Am I missing something here?

3- Re: the fact that it came out in October and US mid-term elections are near.  Well, that's the publication date of the Lancet, and frankly, I dont see how you can use that to disqualify research published by a British medical journal.  Unless  you propose that politically sensitive months in the US electoral calendar should dictate the publication schedule of British medical journals - which is inane. 

Look forward to hearing back about any of these points (or, preferably about the content and methodology of the report itself). 

For sure, this is not good news for the Administration's position that everything is coming up roses is Iraq (and hence they most certainly have a reason to attempt to discredit it), but without scientifically valid criticisms of the data presented, the position that "it's incovenient, hence it's libel" hold no water whatsoever.

Heart,

eu4ea
 
Quote    Reply

EW3       10/11/2006 4:41:53 PM
 
This is the same basic study they did 2 years ago which showed up just before the presidential election.  Maybe that is why I am a bit POed...  From what I have read about this study they used the same statistical method they did that time, which I did read and understand.  And it is very bogus.     
As to it agreeing with other studies, you really need to use more varied sources.  The Lancet is 10x the number of most reports.  This is certainly not a "conservative" source  http://www.iraqbodycount.net/
Most reports are bunched up around 30000-40000, not even close to the lancet.
 
 
 
Quote    Reply

GreyJackal       10/11/2006 5:32:28 PM
Last month alone almost 2700 iraqis were killed in Bagdad alone. Just in the capital.  These deaths are only the ones we know about. The Journalists are mostly reporting while only covering 2-4% of Iraq, the other 90%+ are usually too dangerous(lacking proper security) to do reporting let alone body counting.
 
I think this 655000 figure is pretty accurate.
 
Quote    Reply

EW3       10/11/2006 5:55:53 PM

Last month alone almost 2700 iraqis were killed in Bagdad alone. Just in the capital.  These deaths are only the ones we know about. The Journalists are mostly reporting while only covering 2-4% of Iraq, the other 90%+ are usually too dangerous(lacking proper security) to do reporting let alone body counting.

 

I think this 655000 figure is pretty accurate.


Where do you guys come up with these numbers?  Sources please...
You may want to look at this link: http://www.icasualties.org/oif/IraqiDeaths.aspx
I would also point out that the people doing most of the killing are bad guys.  Either terrorists or dead enders from Sodamns old gang of thugs.  It's not Shia death squads blowing up car bombs in Baghdad.
 
Quote    Reply

EW3    eu4ea   10/11/2006 6:01:25 PM
BTW - the Lancet is published weekly. 
 
In January 2006, it was revealed that data had been fabricated in an article by the cancer researcher Jon Sudbø and 13 co-authors published in The Lancet in October 2005, http://news.bbc.co.uk/2/hi/health/4617372.stm href="http://news.bbc.co.uk/2/hi/health/4617372.stm">[1]. The fabricated article was entitled "Non-steroidal anti-inflammatory drugs and the risk of oral cancer: a nested case-control study". http://www.aftenposten.no/nyheter/iriks/article1199644.ece href="http://www.aftenposten.no/nyheter/iriks/article1199644.ece">[2]. Within a week after this scandal surfaced in the news, the high-impact New England Journal of Medicine published an expression of editorial concern regarding another research paper published on a similar topic in the journal.
Apparently they do make mistakes.
 
 
 
 
Quote    Reply

eu4ea       10/11/2006 6:12:34 PM
I really encourage you to read the report, Platy. 

You may be surprised - it really is an *extremely* well done study, as good as our current understanding of Statistics and Epidemology can make it. As for the results, I couldnt follow the link you sent, (other than to notice that it pointed to iraqbodycount.net.) 

I'd really appreciate it if you could re-send it - and in the meantime I think it's worth pointing out that there's a fundamental difference between other reports published to date and the Lancet's study. 

Most, including both Iraqbodycount, the US military studies, and the UN Assistance Comission for Iraq's studies are 'passive surveilance' studies - that is, they count deaths reported by third parties (such as media reports or morgue records).  This method presents unsurmountable logistical difficulties, particularly in the case of events occuring over a long periods of time and across large land areas - and most particularly in events that occur in areas that are unsafe or otherwise inaccesible.  That difficulty is critical in events like armed conflict, epidemic disease, or natural disasters - indeed, except for very exceptional circumstances, the passive-surveilance approach is not considered a valid method for any of them.

The active-surveilance method (that is, actually goign around counting corpses) presents event greater logistical difficulties and in caotic environments it's unaplicable to anything other than groups that are closed, highly regimented, and highly documented (such as the US military itself.)

To provide some idea of scope of error introduced by passive-surveilance methods, other than Bosnia there are no major conflicts where passive-surveilance methods were able to identify over 20% of the casualties identified through population-survey methods. (see footnotes 10,11,12, and 13 in the Lancet study).

This is the reason why the CIA, the US military, the State Department and the Centers for Disease Control all use population-survey methods when calculating the death toll of events ranging from armed conflict to natural disasters, epidemic diseases and health crises.

Within this context, the Lancet's study is a gem - it's *extremely* well planned, executed, analized and thoroughly peer-reviewed prior to publication.  Pretty much as good as it contemporary Epidemiology can make it - and for good reason.   You can be sure they expected it to be picked apart in minute detail, and they went through crossing every t and dotting every statistical i on it.

Looking past the headline item of "655,00 excess deaths", the most interesting bits are deeper within the study - things like their tracking of trends over time, the impact on different demographic groups (particularly males 15-44), the differences amongst provices, and evolution in the causes of death (violent and not). We'd do well to pay attention to it - there's a lot of valuable information in there.

Futher, the overall result (655,000 deaths works out to 2.5% of the population dead over 3 years of armed conflict) is fully consistent with trends we have observed in other recent wars (Bosnia, Algeria, the Congo, Eat Timor, Dafur, Vietnam, etc). Completely standard - if anything, it's a bit on the low side when compared to other modern conflicts.

What would be *absolutely* unusual would be a death toll on the level of what Bush claimed (30,000 deaths works out to 0.125% of the population dead over 3 years of armed conflict).  That's inane, a *complete* fabrication.  Right up there with communist-era claims that 97% of the population voted for the "Dear Leader" du jour. 

In my view the honest thing to do is to call him on it, rather than to inmediatelly assume bias in a peer-reviewed scientific paper published by a top-notch British medical journal.

Heart,

eu4ea


This is the same basic study they did 2 years ago which showed up just before the presidential election.  Maybe that is why I am a bit POed...  From what I have read about this study they used the same statistical method they did that time, which I did read and understand.  And it is very bogus.     

As to it agreeing with other studies, you really need to use more varied sources.  The Lancet is 10x the number of most reports.  This is certainly not a "conservative" source  http://www.iraqbodycount.net/</div>;" target="_blank">link
Most reports are bunched up around 30000-40000, not even close to the lancet.
      
 
Quote    Reply

eu4ea       10/11/2006 6:28:37 PM
They most certainly make mistakes, Platy - everyone does.

They had a controvesy over that article regarding non-steroidal anti inflamatory drugs on oral cancer.  They also had one over a paper they published in 1998 about possible statistical links between the MMR vaccine and autism.

However - take that, take the number of complex scientific papers they publish every year, and take the level of professional scrutiny to which they are subjected, and what you get is about as close to "infalible" as it's humanly possible to make a journal.

Seriously, the Lancet is no broadsheet - in fact it's probably amongst the top 4 or 5 publications in the planet in terms of scientific accuracy.

eu4ea

BTW - the Lancet is published weekly. 

From:http://en.wikipedia.org/wiki/the_lancet</div>;" target="_blank">link
 

In January 2006, it was revealed that data had been fabricated in an article by the cancer researcher Jon Sudbø and 13 co-authors published in The Lancet in October 2005, http://news.bbc.co.uk/2/hi/health/4617372.stm" target="_blank">link href="http://news.bbc.co.uk/2/hi/health/4617372.stm">[1]. The fabricated article was entitled "Non-steroidal anti-inflammatory drugs and the risk of oral cancer: a nested case-control study". http://www.aftenposten.no/nyheter/iriks/article1199644.ece" target="_blank">link href="http://www.aftenposten.no/nyheter/iriks/article1199644.ece">[2]. Within a week after this scandal surfaced in the news, the high-impact New England Journal of Medicine published an expression of editorial concern regarding another research paper published on a similar topic in the journal.

Apparently they do make mistakes.

 
Quote    Reply

eu4ea       10/11/2006 6:31:40 PM
Oops, I've been calling you "Platy", EW3.

Sorry about that - in the new format I cant see who wrote a post while I'm replying to it, and got mixed-up.
My bad.

Heart,

eu4ea

 
Quote    Reply

Pseudonym       10/11/2006 6:36:04 PM
Isn't this the journal that had the Iraqi households self-report however many casualties they wanted with absolutely no verification?

Funny how that report came out just in time for an election, and this one comes out just in time for an election.

Don't you think if over 600,000 more people had died then reported in the most televised war in history that someone, ANYONE, might have noticed?

Oh no, let's go with the numbers from Lancet.

LOL.

 
Quote    Reply
1 2



 Latest
 News
 
 Most
 Read
 
 Most
 Commented
 Hot
 Topics