Breakdown of violent deaths in Lancet study

tim, I don't complain when you mention the turkey -- I just think your obsession with it is funny. And I have 419 posts on John Lott, so when I mock someone else's obsession with a topic, I'm sending myself up as well.

Aren't you the guy who complains every time I mention a certain fake turkey, or reach double-figures in one of your comment threads? How much longer will you persist with your Lancet obsession? Move on, Lancet boy!

Mike, as I suggested on the other thread, it was correct to delete Fallujah from the analysis of the overall risk ratio as an outlier, but that doesn't mean you can entirely ignore it. Fallujah (along with Sadr City, Najaf, Sammarra and Rammalla) was a high-violence cluster. These high-violence clusters are likely to account for a significant proportion of the overall excess deaths. Within the high-violence clusters, aerial munitions apparently account for a very high proportion of the deaths by violence. So I think it's flawed to argue from the ex-Fallujah numbers exclusively.
In general, however, I think you have a point; this sort of discussion is exactly the sort of reason why I have never been keen on the extrapolated 100,000 number and have always profoundly wished that the debate was being conducted in terms of the overall risk ratio.

Tim:

I don't see how the numbers now work within the framework of the authors' presentation of their results. I'll repeat what I believe is the defining quote from the authors:

" Making conservative assumptions, we think that about 100,000 excess deaths, or more have happened since the 2003 invasion of Iraq. Violence accounted for most of the excess deaths and air strikes from coalition forces accounted for most violent deaths."

60,000 is the number that has commonly been put forward by yourself and others during this debate for the number of violent deaths contained in the excess death toll of 100,000. From the authors' own premise above, we have to be looking at more than 30,000 coalition caused bombing deaths outside of Falluja. How many more, who knows, but the authors clearly state they are the majority of the violent deaths. But we now know that bombing deaths outside of Falluja accounted for only 28.5% (6 of 21) of the violent deaths. As a precentage of 60,000, this works out to only 17,000 deaths.

I may be missing something, but as I said before, the numbers don't work. The authors claim coalition bombing is the number one contributor to their excess 100,000 death toll, but it clearly is not. As I mentioned in an earlier post, the bombing deaths only manage to outnumber violent deaths of "unknown origin" by a factor of 3 to 1. The number of bombing deaths is less than the increase in the number of deaths from heart attacks, it's less than the number of deaths from homicide, and it's less (by a factor of 2 to 1) than the violent deaths that cannot be contributed to the coalition.

As I also mentioned in an earlier post, the figure of 21 violent deaths outside Falluja is only as credible and reliable as the sum of its parts. I believe issues concerning the parts raise many questions about the reliability of the whole.

you quite rightly note that, as a result of the clarification from Dr. Roberts, there is now very little room in the authors' conclusions for any allowance for insurgent deaths. We have 2 " maybes," out of 21, and 1 of them we have no idea whether he might or might not have been an insurgent. Clearly, even 2 insurgent deaths out of 21 violent deaths will cause a vastly lower nationwide extrapolation than the figure of 25,000 insurgent deaths that both you and I have tentatively put forward.

I realize this would have the effect of implying that the 100,000 excess death toll outside Falluja is far too low, but that means we have to assume that the 100,000 figure is reasonably accurate for every other type and cause of death, when in fact the gross under-estimation of the insurgent death toll is yet another indicator that the parts making up the sum have some serious problems, and do not coincide with our understanding of events on the ground.

It should be noted- seeing as no one who comments on this study seems to know anything at all about air dropped / launched weapons (including the authors of the study itself)- that bombing usually creates several wounded casualties for every death. For example, one of the more conservative estimates on german casualties during the strategic bombing offensive in WW2 estimated 300,000 deaths and 800,000 wounded. This ratio of 2-3 wounded for every death is not untypical in warfare.

Therefore, if we assume, as some have tried to do, that nearly all of the deaths where caused by bombing, we would expect to have seen something like 200,000-300,000 wounded. I'd be curious to know if there is any evidence of this... have we really been seeing over 400 casualties every single day? If we haven't then I wouldn't be very confident in any claim saying aerial bombing is responsible.

Wilbur, I don't think I've seen anyone claim that 100,000 deaths were caused by bombing and I don't think you have either.

So overall one can see that the non-violent mortality rate after the invasion was 1.2 times what it was before (69 versus 57, so relatively small error bars), and the violent non-coalition non-Falluja mortality rate was 10 times what it was before (12 versus 1.25, so enormous error bars). About what one might expect from a country whose public health system and security services have been obliterated.
Overall the non-coalition mortality rate was up 40% (81 vs 58).

Hi;

I've greatly enjoyed reading the blog. Well done.

I've also read the paper, and know a little (but obviously not enough) about statistics. Could someone explain what this line (from page 2 of the paper) means.

"We assigned 33 clusters to Governorates via systematic equal-step sampling from a randomly selected start."

I have looked for an answer on the 'net but have been unable to find one. How does this method differ from randomly assigning clusters to Governorates based on the Governorates population sizes?

Secondly; part of the sampling procedure involves: (1) drawing a map, (2) generating a random easting and northing, (3) going to the nearest 30 houses. It seems to me that this method is quite seriously flawed - though I admit I can't think of any improvement to it off the top of my head.

The idea of sampling is to choose the population studied at random. The study's technique (partly) selects a geographical location at random, not a chunk of population. This will systematically over-sample the population in areas of low population density and under-sample the population in areas of high population density. There are going to be variations in population density, because that's how human geography works. If there is any difference in mortality between high and low density areas, this is going to throw the results of the study. Even a small difference in mortality between high/low density areas could have a large result on the extrapolated results.

Thoughts?

I think the study is a very valuable contribution, and have no wish to rubbish it or see it rubbished. But the above seems to me a valid and substantial critique of its methods that I haven't seen anyone so much as mention.

D Squared:

I'll have to be brief, since I'm bagged, and ready for bed. I'll be more detailed tomorrow.

As I mentioned in my latest post in the earlier thread, Falluja was excluded from the 100,000 estimate because the authors weren't prepared to try and sell an excess death figure of 300,000.

To some degree, this debate is about how high an excess death toll one is willing to try to defend. Are you willing to say that 200,000 is plausible, based not just on statistical arguments, but on an objective consideration of the media reports and other information that provides us with a picture of the situation on the ground? How about 300,000? Do we dare go higher? I ask this, because I believe this study's methodology, if replicated, is capable of providing us with figures of 300,000 or higher.

In fact, we know this is true, because the actual Lancet study did just that. If we accept the data as a whole, without excluding Falluja, we are left with a nation wide excess death extrapolation of 300,000.

So D Squared, when you suggest that the near absence of data from Sadr City, Najaf, Ramadi, etc actually could compensate for the death numbers excluded from Falluja, you're making an implied case for the very death toll the authors didn't want to deal with, and one which blows the lid off the confidence interval they adopted.

Wilbur:

I think you have raised an interesting sidebar to the subset of bombing deaths. I can't give you anything, anecdotal or otherwise, that demonstrates an absence of, or evidence of, a flood of bombing wounded in the range of say 100,000 to hospitals across Iraq.

Enquiring mind:

I think you raise some relevant issues. I see where you're going with them, and I think they do add yet another element of uncertainty to the study results, but I'm not a statistician. I've seen them touched on elsewhere. Dr. Roberts addressed the issue of the effects of population density on the accuracy of the results. I can't recall if it was in the study, the Medialens e-mails, or another interview I read online, and frankly I'm too tired to look tonight. I do recall it was definitely a concern to them, particularly as it related to the possibility that population shifts had occurred as a result of the conflict.

Tim, D Squared or Scott are the ones who could provide the expertise to give you an indication of the statistical consequences of population density fluctuations and errors.

Tim Lambert writes, "In other words, 95% or more of the 100,000 excess deaths were civilians, so it is not wrong to describe the findings as 'about 100,000 civilian deaths'."

Let me get this straight...you're saying that out of the 100,000 "excess deaths" projected from the Lancet study, 95,000+ were civilians (i.e., only <5,000 were combatants)?

If that is indeed what you're saying, I'm curious about your estimated breakdown for that 95,000+:

What percentage of those 95,000+ "excess" civilian deaths were violent deaths directly caused by Coalition forces?

What percentage were violent deaths caused by people/groups other than Coalition forces?

What percentage were non-violent deaths?

So only 4000 combatants were killed in the war, and 96,000 were civilians? I think that little pop I just heard was your credibility going. Even the authors of the study said in their latest article that the reason for military scepticism was that the media misreported women and children as being the main victims of airstrikes. And, they said, the majority could have been combatants. This nonsense about 95% being civilians only finishes off the last bits of credibility this study had. It's over.

Do you think the relatives of these combatants are going to say that their sons was killed in a fair fight? Or do you think it likely they will make up some crap about being shot at a check point etc, as people in war always do? Come on, people! Do you have absolutely no understanding of the political and cultural situation in Iraq at the moment? This is absurd.

Brian, insurgent deaths are likely to be concealed altogether from the surveyers. If the survey undercounts them it doesn't follow that there is anything at all wrong with the rest of the data. I think that the Iraqis who did the surveying in Iraq have way way more understanding of the political and cultural situation there than you do.

Mike: I've mentioned before that I don't like the extrapolated number. However, I would say that the true death rate, based on the evidence, is more likely to be in the top half of the confidence interval than the lower half, and that the authors weren't blowing smoke when they called the 1.5x figure "conservative". I'd say that the effect of the war has been to lower the death rate in the Kurdish North, to raise it by up to 50% in the majority of the country and to kill a lot of people in the high-violence clusters. Beyond that, I think I'd refuse to be drawn; unlike the Lancet team I'm not that interested in the 48-point headline number.

"insurgent deaths are likely to be concealed". Oh right, so for weeks when anyone has questioned the integrity of Iraqis who took part in this study, we have been told that we are being totally ridiculous. But now suddenly Iraqis are all "concealing" insurgent deaths? Get real. It is far more likely that Iraqis would lie about the cause of the death, rather than pretend the person didn't exist. You must have seen Iraqis at Falluja claiming that dead men found with guns in their hands were civilians. Iraqis are always on hand to claim any incident where US troops kill someone was a civilian. No one ever owns up to being a terrorist. That is the political reality.

Hi, Tim:
>
I think there's been a technical glitch at your site, as this post seems to have been...misplaced. ;-)

From your response to Brian, it seems even more clear to me that you *are* claiming that 95,000+ of the 100,000 "excess" deaths outside Fallujah were non-combatants.

Given this...how do I put this delicately?...ludicrous claim, I really would like to see a further breakdown of that 95,000+ civilian deaths. Specifically, how many of 95,000+ civilians were:

1) Violent deaths caused by Coalition forces?

2) Violent deaths caused by non-Coalition folks?

3) Non-violent deaths?

Mark

P.S. I guess I should add a fourth category of "Other," since your clever (aka, fevered) mind may have come up with a fourth category that's outside those three. ;-)

Hi, Tim:

Your old friend Mark here again. :-)

From your response to Brian, it seems even more clear to me that you *are* claiming that 95,000+ of the 100,000 "excess" deaths outside Fallujah were non-combatants.

Given this...how do I put this delicately?...ludicrous claim, I really would like to see a further breakdown of that 95,000+ civilian deaths. Specifically, how many of 95,000+ civilians were:

1) Violent deaths caused by Coalition forces?

2) Violent deaths caused by non-Coalition people?

3) Non-violent deaths?

Mark

P.S. I guess I should add a fourth category of "Other," in case your clever (fevered?) mind can come up with a fourth category that's outside those three. ;-)

Mike, on the Fallujah cluster I don't think the authors or anyone else believes that that the 52 deaths (out of around 200 or so) means that one quarter of the people in the province containing Fallujah were killed.
That's the 200,000 death toll referred to in the paper, and nobody including the authors is silly enough to believe it. What it does show is that the heavily bombed neighborhoods in Iraq might be experiencing extremely high civilian casualties. There was one heavily bombed neighborhood out of 33 in this survey, and the authors point out that there were many neighborhoods in Fallujah and other places which appeared to have been hit just as hard. I'm not a statistician either, but common sense suggests that this means the non-Fallujah data (which didn't include any heavily bombed neighborhoods even in Sadr City) underestimates the civilian death toll. Common sense suggests that you can't say much more than that--to me the value of the Lancet study comes from suggesting that the overall violent death toll is in the high tens of thousands, maybe more, from a mixture of crime, terrorism and American military action, and that the heavily bombed areas are suffering horrific death tolls. How many neighborhoods are heavily bombed? 1 out of 33 in this study and nobody knows what the actual percentage is. Maybe some enterprising sort could find out from satellite photographs, if they haven't all been classified.

BTW, from a propaganda viewpoint the insurgents are doing the US government a favor by keeping reporters out and kidnapping and killing peace activists like Margaret Hassan. The Fallujah neighborhood was hit hard in April, when it looks like around 10 died, and again in June (I think the bar graph shows 3 or 4) and then really hard in August (around 30) and then again in September (around 5). Were reporters freely going around Fallujah gathering horror stories about families being killed or did they go interview the refugees? Not that I noticed --there were a handful of peace activist types who went into Fallujah in April, and there was that Iraqi AP reporter who said he saw American helicopters
shooting at Iraqi civilians trying to cross the river during the recent assault, but for the most part, Westerners couldn't report on what was happening there in recent months because they'd be killed if they went wandering around on their own. Probably most American reporters are happy just passing along US government propaganda anyway, but the more independent-minded have been prevented from doing first-hand reporting by the actions of the insurgents.

Mike, on the Fallujah cluster I don't think the authors or anyone else believes that that the 52 deaths (out of around 200 or so) means that one quarter of the people in the province containing Fallujah were killed.
That's the 200,000 death toll referred to in the paper, and nobody including the authors is silly enough to believe it. What it does show is that the heavily bombed neighborhoods in Iraq might be experiencing extremely high civilian casualties. There was one heavily bombed neighborhood out of 33 in this survey, and the authors point out that there were many neighborhoods in Fallujah and other places which appeared to have been hit just as hard. I'm not a statistician either, but common sense suggests that this means the non-Fallujah data (which didn't include any heavily bombed neighborhoods even in Sadr City) underestimates the civilian death toll. Common sense suggests that you can't say much more than that--to me the value of the Lancet study comes from suggesting that the overall violent death toll is in the high tens of thousands, maybe more, from a mixture of crime, terrorism and American military action, and that the heavily bombed areas are suffering horrific death tolls. How many neighborhoods are heavily bombed? 1 out of 33 in this study and nobody knows what the actual percentage is. Maybe some enterprising sort could find out from satellite photographs, if they haven't all been classified.

D Squared:

I think there's a conflict inherent in your statement from your previous post:

"I've mentioned before that I don't like the extrapolated number. However, I would say that the true death rate, based on the evidence, is more likely to be in the top half of the confidence interval than the lower half,......"

You've made similiar statements elsewhere concerning your disagreement with the authors extrapolating large numbers from the study data, then you turn around and not only do the same thing, but come in with a higher number than the authors!

If you're " basing " your own projections on " the evidence," (and I assume you're referring to the study data), I submit that you're relying on some very shaky evidence, as I've tried to point out with some of my arguments concerning a finer analysis of the subsets that constitute the death figures.

It isn't only the violent deaths following regime change that have driven up the death rate in Iraq. As you'll recall, just over half (22 to 21) of the excess deaths outside Falluja were non-violent. Approximately 1/3 (7 out of 22) were from an increase in heart attack and stroke deaths, but all 7 were found in the children (2) and men(5) categories. The combination of women and the elderly recorded exactly the same number of such deaths before and after regime change. This increase of 7 (from none pre-war) had a significant impact on the authors' ability to further elevate the mortality rate in post war Iraq. But is this elevation a direct consequence of the war, or simply chance? It seems odd that no men died of heart attack before the war, but 5 after. Similarily, no children died from such causes pre-war, but 2 did after. Were the children's deaths the result of congenital defects, that had nothing to do with the war? Did any of these 7 victims die as a result of inadequate medical care caused by the war, or would they have died without invasion? If we look at the women in realtion to this, 3 died of heart attack before the invasion, only 1 after. (The elderly remained relatively consistent, 8 deaths before, 10 after).

It's difficult to make the argument that the 7 men and children who died from heart attack and stroke can be attributed to invasion, when we have 3 women dying before invasion of the same cause, and only 1 after.

We also have a statistically crucial jump in deaths from accidents, which seems to have a much greater impact on the increased rate of death than the much bandied-about infant mortality figures. 4 people died in accidents before the war, but the number after regime change jumped to 13, for an increase of 9. Incredibly, with all the debate going on over coalition bombs allegedly killing tens of thousands of Iraqis, we've got 20% of the excess deaths outside Falluja coming from accidents. I don't take issue with anyone making a case that instability in post-regime change Iraq creates an environment for more fatal accidents. However, first of all, is it reasonable to blame all of this excess accident toll on the coalition, or is it yet another death subset that is volatile and probne to shifting from study to study? Secondly, the authors make it clear that the number one killer among the 100,000 excess deaths they project is coalition aerial bombardment. Yet deaths from accidents outnumber bombing deaths outside Falluja by a factor of 3 to 2.

The increase in heart attacks and the increase in accident deaths constitute 37% of the excess deaths recorded outside Falluja. It's my belief that such a percentage could dramatically fall in subsequent surveys, and in any case, it seems unfair to blame a 100,000 excess death figure on the coalition when so much of it comes from these two causes of death.

It's also interesting to note that many who have defended the study dismiss the possibility that Saddam was killing tens of thousands of Iraqis in the 2 or 3 years immediately prior to regime change. This also serves to provide a justification for the extremely low violent death rate reported before invasion (only one such death).

While Saddam's victims didn't generally show up in the morgue, and we have no way of accurately determining the level of Saddam's killing immediately before the war, I suspect the Left is right, and that the killings weren't in the tens of thousands.

But there definitely were some. It's ironic that the only violent death attributed to Saddam's agents is recorded in the post-invasion figures, and serves to add to and drive up ever so slightly the excess death figure used to castigate the coalition.

In closing, you'll get no argument from me that the deatth rate in Iraq has definitely risen since Iraq. For many reasons, I see this study as being unable to provide us with a reliable post-war increased death rate calculation, and also unable to accurately apportion the excess deaths into the most germane of the death subsets

Mike;

"It's difficult to make the argument that the 7 men and children who died from heart attack and stroke can be attributed to invasion, when we have 3 women dying before invasion of the same cause, and only 1 after."

No it isn't.

There's considerable evidence that psychological stress can increase the incidence of sudden cardiac death. This is particularly the case very sudden and severe events, such as an earthquake or a war. You can search PubMed for the references yourself, if you're interested.

While I'm writing, does anyone out there have an explaination of the sampling strategy used? As I requested above.

Donald:

I agree with you that the authors did not believe that the 52 deaths in Falluja could be uniformly extrapolated into a 200,000 death toll for the entire cluster.

However, as I've stated elsewhere in this debate, if bombing has occurred in the widespread, highly destructive manner that the authors allege in their study conclusions, then we have many, many neighbourhoods in Iraq with multiple bombing deaths in multiple households. You simply can't get to 30,000 bombing deaths or higher without this.

If this is correct, then additional surveys using the same methodology could easily approach Falluja-like death tolls from a combination of 2 or 3 of such clusters reporting say, 10-15 deaths each. These clusters would not be excluded as outliers, and before you know it, you've got near-Falluja bombing numbers, without Falluja, and then the authors WOULD extrapolate an excess death toll approaching 300,000.

It's worth repeating that the tens of thousands of bombing deaths projected to have occurred means we should have many hundreds of neighbourhoods reporting multiple deaths in multiple households. The study can't give us 1 such neighbourhood outside Falluja, which the authors repeatedly claim to have ignored for purposes of the 100,000 excess death toll.

If, as you suggest, "..... that the heavily bombed areas are suffering horrific death tolls.", this study isn't providing us with evidence of that. We have evidence of a single heavily bombed area. Coincidentally, it's also the one area we know has been heavily bombed from media reports.

This study hasn't given us much insight, in my opinion, concerning the distribution of bombing deaths across Iraq. As I pointed out at the beginning of this thread, the authors tell us they believe at least 30,000 have died from bombing outside Falluja. I haven't seen any convincing evidence of this from their own data, in fact, I see more that contradicts this than corroborates it.

The paper authors mention that there were several other heavily damaged cities like Ramadi, Najaf, and Tallafer which were not selected. They also add that Sadr City was selected, but they happened to pick a neighborhood with no deaths from all the fighting that's been going on. So there may be a fair number of neighborhoods that have suffered heavy bombing.
It boils down to this--by chance, they found one cluster out of 33 with a horrific death toll. If it's really true that 1 out of 33 neighborhoods in Iraq lost around 1/4 of their people to bombs, you'd have 200,000 deaths added to the 98,000 (with huge error bars) of the rest of the study. 200,000 bombing deaths is hard to believe, so the true ratio is presumably lower. But suppose it is ten times lower-1 out of 330. Then that would still be 20,000 deaths from air attack to be added onto the 98,000 from various causes. And that is leaning over backwards to be conservative--if the correct ratio is really 1 out of 330, then it was unlikely that in a sample of 33 they'd stumble across such a neighborhood, which suggest that the extra death toll from heavily bombed neighborhoods is greater than 20,000. But anyway, I'd say the study makes a good case that many tens of thousands of people have died violently and also shows that there's a good chance many of them died from American bombing. Someone needs to repeat the study with more samples, or alternatively an actual count needs to be done. Better yet, Americans shouldn't flatten any urban areas from now on. That's a bizarre way to win hearts and minds, whatever one thinks of the war.

Enquiring Mind:

I think you're stretching things a bit here. As I pointed out earlier, you had a jump from zero to 7 among men and children, but the deaths from heart attack fell from 3 to 1 for women. Isn't it more likely that the 2 children died from congenital problems? One would think the elderly would be the demographic most susceptible to stress-induced coronaries, yet we see only a modest increase in heart attacks among the elderly (from 8 before regime change up to 10 afterward).

Would it be that unexpected to find a second survey revealing data that showed less men dying of heart attacks AFTER invasion than before? Zero heart attacks among a sampling of 2200 adult males seems a bit of a fluke.

Without knowing the medical histories behind each of the 7 cases among the men and children, and in light of the contradictory numbers from the women and elderly, I'm not sure your stress-related assumption can be given that much weight.

It's also noteworthy that the post-invasion population that made up the sampling was 430 persons higher than the pre-invasion sampling (7868 versus 7438), and the post-invasion time frame is 3 months longer than the pre-invasion time frame (17 1/2 months against 14 1/2 months). For purposes of the raw numbers, this should have some small bearing on deaths from non-violent causes, since you have more people involved, and over a longer time frame, than the pre-war totals.

Donald:

You've confused the figures somewhat with this statement:

"Then that would still be 20,000 deaths from air attack to be added onto the 98,000 from various causes.",p>

The 98,000 excess death estimate from the authors includes, by the consensus of most who have debated this study here, more than 30,000 bombing deaths. You can't tack 20,000 onto the 98,000 estimate. Your new estimate of 20,000 bombing deaths would replace the 30,000 component, and would in effect reduce the estimate by 10,000, not increase it by 20,000.

"But suppose it is ten times lower-1 out of 330. Then that would still be 20,000 deaths from air attack to be added onto the 98,000 from various causes. And that is leaning over backwards to be conservative--if the correct ratio is really 1 out of 330, then it was unlikely that in a sample of 33 they'd stumble across such a neighborhood, which suggest that the extra death toll from heavily bombed neighborhoods is greater than 20,000."p>

Donald, what you're really doing here now is throwing out a guesstimate, and in doing so, acknowledging that the study itself isn't giving us the data required for a reasonably accurate accounting of the death toll from bombing.

This problem is further illustrated by the treatment of the Falluja data. It does extrapolate to 200,000 deaths, all from bombing. Obviously, this doesn't appear plausible in the least. But how would we make a plausible estimate from the Falluja data? Where would you peg the deaths at in Falluja? I don't know how you could make a confident projection. IBC announced with great fanfare their estimate that 600 civilians had died in the April heavy fighting there. The truth is, no one has any idea what a fairly accurate figure would look like for Falluja casualties. You could make a case for almost any number, from 1,000 to 100,000, higher if you want. The authors did not provide any estimate of their own for a total number in Falluja that might be defendable. The study doesn't allow them to.

My question then is, if Falluja is an open number, where did the authors find the confidence to extrapolate a fairly firm number for bombing deaths from the rest of Iraq, based on only 6 deaths in 4 clusters?

As I've said several times before, the bombing pattern by the coalition does not appear to be carpet bombing of large urban centres, with the exception of Falluja. You mention that Ramadi and several other cities were also extensively damaged, but not sampled. Correct me if I'm wrong, but I don't believe the destruction in these urban areas is comparable to Falluja.

In addition, destruction of sections of built up areas does not uniformly equate to large casualty figures. For this to have occurred (consistently heavy casualties wherever we find areas of heavy damage) we have to assume that the population decided to stick around to have the bombs fall on them. Obviously, in some cases Iraqis would have had no opportunity to avoid such a fate, but I believe it's a faulty assumption to assume this has occurred in every instance of heavy bombing damage in urban areas.

Donald, if there are 20,000 bombing deaths, as you estimate, or more than 30,000, as the authors claim, then we've got hundreds of neighbourhoods with multiple deaths in multiple households, as I've stated before. Let's assume a second sampling using identical methodology found just one such neighbourhood outside Falluja, and say 15 deaths were reported as a result. The authors were willing to extrapolate more than 30,000 bombing deaths from 6 non-Falluja deaths. How high would the extrapolation be from 15? Let's say your 20,000 estimate was very close to being right. Would not a survey finding 15 bombing deaths among 7,800 people extrapolate to a nation-wide figure that is 3 time higher or more than the actual figure of 20,000?

To illustrate this further, you could have a true death toll from bombing of 5,000 (I'm not adopting this, I believe it's inaccurate and low). A cluster sampling could still wind up taking in a single neighbourhood with multiple deaths in multiple households from bombing, resulting in a nationwide extrapolation that is hugely skewed, and many times higher than the true toll.

Enquiring Mind:

You accuse me of speculation, but aren't you taking speculation to an entirely new level with this statement?:

"It's entirely possible that deaths rose among men and children because these two groups had stress elevated most by the war, and it's entirely possible that children may have congenital problems but these only kill them in certain circumstances."

Is it realistic to suggest that children and men are affected by the stress of war more than women or the elderly? Do you have any evidence to substantiate that? What possible means of measure could there be to even attempt to prove that?

"You then engage in speculation about the causes of death. You have not cited any evidence that your ideas are based upon."

Enquiring Mind, I can't very well cite evidence that the study doesn't give us. We don't know the medical backgrounds of the 7 men and children who died after regime change. We don't know if they died from a denial of health care services, if they were very ill from heart disease prior to invasion, etc. These are relevant factors that could confirm whether these deaths can be remotely attributed to the war itself or not.

The fact is, the elderly are more physiologically prone to death from heart attack than any other demographic. We don't see a dramatic increase in deaths from heart attack among the elderly. We see a drop from 3 to 1 among women. My point remains, these two numbers make it difficult to claim from this particular study data that the war is causing an increase in heart attack deaths.

I'm not disputing your claim that there are studies which point to war causing an increase in heart attack mortality rates. I am stating however, that I don't see this study providing proof that the excess deaths from heart attack and stroke can be pinned on the war.

Donald Johnson wrote: "Mike, on the Fallujah cluster I don't think the authors or anyone else believes that that the 52 deaths (out of around 200 or so) means that one quarter of the people in the province containing Fallujah were killed. That's the 200,000 death toll referred to in the paper, and nobody including the authors is silly enough to believe it."
The whole paper relies upon cluster sampling. If this method doesn't work in Falluja, why should it work elsewhere ? There is even less integrity in discarding part of your data, because you don't like the results, and only using the part you do like.
j

Interesting debate. I'd heard about it earlier, of course, but this is the most detail I've seen.
I get the impression of a pyramid standing on its point. Huge conclusions based on tiny samples and arguable extrapolations.
One commenter hit the jackpot in mentioning the likelihood that civilians would leave areas in which terrorists were beginning to show up.
That appears to have been the case in Fallujah, with some reports suggesting more than 95% of the civilians were gone by the time the fighting started.
But, for the fun of it, let's take 100,000 as the figure. There is clearly a blame component implicit, if not explicit, and it's all pointed at the US. If terrorists kill civilians, does that fall the US' share? According to the study, it does, since we're talking about increase in deaths beginning with the US invasion.
Ostensibly, the study simply counts numbers. Clearly, it's designed to blame the US for all of them and is used so by a good many people.
How much fun would it be to have a study which said, for example, "Ten thousand terrorists killed in combat with Americans, twenty thousand civilians killed by terrorists, twenty thousand civilians killed by Americans"?
Obviously, that wouldn't be much use at all.
But that, or whatever the numbers are, is what actually happened.
It's the US' fault that they made the terrorists so mad that the terrorists started killing civilians.
If the US troops weren't such tough targets, this wouldn't have happened. Darn Americans, anyway.

"But anyway, I'd say the study makes a good case that many tens of thousands of people have died violently and also shows that there's a good chance many of them died from American bombing."

I wonder if The Lancet had an article that looked at mortality in Germany in the 17 months after D-day compared with the 14 months before?

My guess is that they didn't. Politics were different back then.

Golly, Mark. Brilliant. I'm quite sure you're trying to bait someone into disagreeing with you because you have digits. Right? Mark always backs his statements. Some sort of R^2 showing correlation between Germany and Eye-rack. T-stat indicating such. You vernmintul injuneers always have the digits. Spit 'em out.

D

Dano, please do not feed the troll.

Mike;

I think you're stretching things a bit here. ...you had a jump from zero to 7 among men and children, but the deaths from heart attack fell from 3 to 1 for women. Isn't it more likely that the 2 children died from congenital problems? One would think the elderly would be the demographic most susceptible to stress-induced coronaries, yet we see only a modest increase in heart attacks among the elderly...

I'm not stretching things. You suggested it was incredible that the War could have caused increased deaths from heart attacks. I pointed out that there's a large body of evidence that suggests this is exactly what wars do.

You then engage in speculation about the causes of death. You have not cited any evidence that your ideas are based upon. It's entirely possible that deaths rose among men and children because these two groups had stress elevated most by the war, and it's entirely possible that children may have congenital problems but these only kill them in certain circumstances.

Would it be that unexpected to find a second survey revealing data that showed less men dying of heart attacks AFTER invasion than before? Zero heart attacks among a sampling of 2200 adult males seems a bit of a fluke.

No it isn't a fluke. There were only 4 deaths among adult men reported in the period studied. It's quite possible that no one died of that cause in the period studied.

Without knowing the medical histories behind each of the 7 cases among the men and children, and in light of the contradictory numbers from the women and elderly, I'm not sure your stress-related assumption can be given that much weight.

Again, it isn't an assumption. The same thing has been seen to happen in similar situations. The picture is consistant with what we know from elsewhere. If you think the data suggests otherwise, I suggest you use a probability model to show that the data is unexpected, rather than asking rhetorical questions about what you guess it shows.

Dano writes, "I'm quite sure you're trying to bait someone into disagreeing with you because you have digits. Right?"

I'm not trying to get anyone to disagree with me, Dano. I think we'd all expect that a mortality study of Germany in the 17 months after D-day would show "excess mortalities" compared to the 14 months before D-day. Wouldn't you expect that to be true?

"You vernmintul injuneers..."

Some more of Dano's classic grade school name-calling. Well, I guess if one can't debate facts (and one's mentality has never advanced beyond grade school!) that's what's left. But it does get tedious.

Tim Lambert adds, "Dano, please do not feed the troll."

The Tim Lambert definition of "troll:" "Someone who asks uncomfortable questions."

BTW, Tim, I'm still interested in the breakdown of your estimate of 95,000+ "excess civilian deaths" in Iraq. How many of them were killed by Coalition forces? How many were killed by non-Coalition people? How many died from non-violent causes?

Mike;

I'm not sure that this is getting either of us anywhere. I think I've established that your original point that there is no way an increase in heart attacks could be pinned on the War is nonsense, which is what I intended to do. In fact I think there's a strong case that the increase was caused by the War - after all that's what the data here shows and that's what has happened elsewhere.

But I don't think there's anything I can say which will change your mind.

My point remains, these two numbers make it difficult to claim from this particular study data that the war is causing an increase in heart attack deaths.

There was an increase in heart attack deaths post war, the data shows that much. It just wasn't consistant across all groups - nothing that I said suggested it had to be. I can't replicate the study with the data at hand but, based on the data they show, it is unlikely that that heart attack deaths after the war could have been generated by the risk given by data before the war.

The fact is, the elderly are more physiologically prone to death from heart attack than any other demographic. We don't see a dramatic increase in deaths from heart attack among the elderly.

The reason for this is blindingly obvious. There is only a small number of people in that group and they were dropping like flies from that cause before the war (1/33/year). The additional risk produced by the War is slight, but it's multiplied across a large group for men and children and only a small group for the elderly, who already had a high mortality from that cause in the first place.

It is entirely plausable that the war caused an increase in heart attack deaths.

It isn't necessary to postulate a stress-related increase in heart attacks in order to explain the increase in mortality from this cause - the collapse of the medical infrastructure means that a higher proportion of heart attacks probably result in death.

I think you misunderstood my argument, Mike. The 98,000 figure in the Lancet study is based on the non-Fallujah data. If you relied entirely on the non-Fallujah data, you get these deaths with an uncertain breakdown between deaths caused by insurgents, criminals, and Americans, along with increased infant mortality. Some of those deaths were apparently caused by American airpower (6 out of 21), apparently from the kinds of incidents you read about from time to time, such as when the cameraman got killed by an American helicopter.

There's a large uncertainty in this 98,000 number, of course, and an even larger uncertainty in the breakdown of those deaths. In these 32 neighborhoods, there were none that suffered from heavy American bombing, so the 98,000 (plus or minus 90,000) number wouldn't refer to deaths from such attacks. We're talking about murders and stray bullets and rockets and mortars here, plus increased infant mortality.

Okay, then there is the Fallujah outlier. This suggests (doesn't prove) that there may be a large number of neighborhoods which were hit very hard by American bombs. This neighborhood lost about a quarter of its population and one such neighborhood turns up in a random sample of 33 neighborhoods. So the authors (they're explicit about this, but I don't have my copy handy) think that it is likely that 1 or 2 such neighborhoods would crop up in most random selections of 33 neighborhoods. So the deaths that come from such attacks aren't part of the 98,000 deaths cited above. But what the authors do (if my memory is correct) is suggest that the true death toll is therefore likely to be in excess of 98,000 due to this added death toll from severely bombed neighborhoods. Unfortunately the sample size was only 33 and only one such neighborhood appears, so it's hard to say how common such neighborhoods are. The most likely estimate (if you relied solely on the Lancet study) would be 1 out of 33, and that would give you a ballpark figure of 200,000 extra deaths to be added to the previous 98,000. (Ignoring the error bars). I doubt this and I think the authors doubt it. So what I suggest is that the Fallujah outlier is a fluke,either in its total death toll (maybe most heavily bombed places lose a smaller fraction of the population) or in the way it popped up in a sample of only 33. I suggested using the same death toll and a ten times lower frequency, not because I think this is right, but because it is fairly conservative. In that case you'd get 20,000 deaths in Iraq from neighborhoods that were heavily bombed, on top of the 98,000 deaths that one calculates by completely ignoring the Fallujah outlier.

My particular numerical assumptions are arbitrary, but you do have to ask yourself how probable it is that a severely bombed neighborhood would appear in a random sample of 33 if they were in fact extremely rare. So I'd tentatively suggest that there may be at least tens of thousands of deaths in severely bombed neighborhoods, which is what my 20,000 figure was meant to illustrate. (I don't mean it as a literal figure). I think the authors of the Lancet paper were reasoning in a somewhat similar way. And you're right--this would mean many hundreds of Lancet size neighborhoods had been bombed, where a neighborhood means a cluster of 30 households. Is it possible that this many places could have been hit causing tens of thousands of deaths and even more wounded? I'm not sure.

Enquiring mind:

You're putting words in my mouth when you state "I think I've established that your original point that there is no way an increase in heart attacks could be pinned on the War is nonsense, which is what I intended to do."

I didn't say there was " no way " the war was responsible for the increase. Have another look at what I've posted. I've said I don't see this study providing proof that its numbers are representative of heart attack deaths for all of Iraq, and I also don't consider the increased deaths among the sample as being conclusively linked to the war.

"In fact I think there's a strong case that the increase was caused by the War - after all that's what the data here shows and that's what has happened elsewhere."

The data shows there was an increase in heart attack and stroke fatalities among the 7,800 people that made up the sample, but you're trying to pass this off as an accurate extrapolation for the entire country. In effect, you're saying, " I like the numbers from this particular sample, they assist my argument, so therefore they're proof that the ratio holds true for all of Iraq."

The problem is, they can't be representative for purposes of extrapolation. The study recorded zero deaths among adult males for the 14 months prior to regime change. It isn't possible, with a population of 24 million people, that no adult males died of heart attack during that time. Obviously, thousands of men have died during that period. How can you measure an increase, if there is one, when the data provides a zero value that we know can't hold? The same applies to the figures for children. We have 2 post war deaths, but zero prior to regime change. The zero is useless for extrapolating. Finally, we have women dying at a rate of 3 to 1 BEFORE regime change. What are the chances that this is accurate? It makes no sense, and conflicts with your assumption that fatal heart attacks have risen dramatically, unless we accept the premise you put forward earlier, that wartime stress affects children and men more than women and the elderly. Are you sure you want to go there?

I'm not discounting there could be an increase in heart attack mortality, and that some of it, possibly even all of it, could be caused by the war. But the study shows a 61% increase in heart attack mortality for the entire sampling. We have no idea, from this study alone, if this is accurate for the entire country. With numbers as small as we're working with among the excess death toll, 7 additional heart attack deaths has a real bearing on the 100,000 excess death toll the authors claim. As I've stated before, it's 1 more than the non-Falluja bombing deaths, which the authors cite as the primary cause of death for their 100,000 estimate. Additional studies would provide a clearer picture, and the actual increase, if any, could be much lower.

Ian and Donald;

I've no time to reply tonight, due to work, hopefully tomorrow.

Dear Donald
your last post is a perfect study. You make numerous assumptions, false statements, and generally chuck what little data there is out the window, before coming to your preordained conclusion.
If you read my last post, you could start to engage with the data.
j

Tim Lambert wrote:
Dano, please do not feed the troll.
Anyone who disagrees= troll
from there, it is perfectly easy to understand why Tim doesn't see any evidence that contradicts his views. No evidence has been provided that undermines the Lancet study- except from trolls, and trolls don't count.
Tim even gets to make stuff up !
j

In all of this banter, I find the denialists like alcoholics who deny their affliction: "The opposite of everything is true". There's no doubt whatsoever that the combined effects of Gulf War I, when the civilian infrastructure of Iraq was deliberately targeted by coalition bombs, continued bombing on an almost daily basis over the next 13 years, crippling sanctions, and then a final invasion (Gulf War II) have killed hundreds of thousands, and perhaps over a million Iraqis. Much of the country has been bombed to smithereens. Ask Denis Haliday or Hans von Sponeck, both senior U.N. officials who oversaw the distribution of humanitarian aid to iraq during the sanctions regime. Both resigned over what they described as "western genocidal policy". But to those in denial, our governments are virtuous defenders of democracy, human rights and freedom. Any attempt to dent this veneer is met with unrepentent denial and angry ripostes, most of which have little substance, other than to defend an illegal invasion which violated the Nuremburg code and dozens of other international laws. There are countless other examples of other atrocities in which western governments went into "alcoholic denial" soon after the carnage was committed. One of the best examples was the U.S. subjigation of the Philippines from Spain in 1901-02, when up to a million Philippinos were slaughtered by U.S. forces. One U.S. general claimed that he "Wants to make a howling wilderness in the country"; one in 6 on the island of Luzon was massacred. Yet, within a year, U.S. historians were going into their denial mode, claiming that the U.S. is a "Country of laws, not of men", and asserting its basic benevolence as a standard bearer of international law. Nothing has changed. Iraq was an imperial conquest, in which civilians were and are "unpeople", but our propoganda machine (the media) has normalized and sanitized the many atrocities committed in our name. Many of the contributors to this blog continue in the rich tradition of western denial and complicity. History, as with the other examples, will vanquish these arguments. Thanks to Tim for his efforts in support of the Lancet study.

For those in denial, this interview with non-embedded reporter Damar Jamail sums up the situation in Iraq. I would suggest those here (Mike, Richard, Brian etc.) desperate to support U.S. actions have a good read: http://www.zmag.org/content/showarticle.cfm?SectionID=15&ItemID=6857

This letter from the authors of the study defending their findings was published December 12 in the UK newspaper The Independent on Sunday:Iraq toll: at least 100,000 You reported ("Polish hostage held in iraq is released unharmed", 21 November) the Foreign Secretary's response to our study published in The Lancet of civilian deaths in Iraq. It is heartening that Jack Straw has addressed the topic in such detail. However, his response includes an apparent misreading of our results.Our study found that violence was widespread and up 58-fold after the invasion; that from 32 of the neighbourhoods we visited we estimated 98,000 excess deaths; and that from the sample of the most war-torn communities represented by 30 households in Fallujah more people had probably died than in all of the rest of the country combined.Fallujah is the only insight into those cities experiencing extreme violence (ie Ramadi, Tallafar, Fallujah, Najaf); all the others were passed over in our sample by random chance. If the Fallujah duster is representative, there were about 200,000 excess deaths above the 98,000.Perhaps Fallujah is so unique that it represents only Fallujah, implying that it represents only 50-70,000 additional deaths. There is a tiny chance that the neighborhood we visited in Fallujah was worse than the average experience, and only corresponds with a couple of tens of thousands of deaths. We also explain why, given study limitations, our estimate is likely to be low. Therefore, when taken in total, we concluded that the civilian death toll was at least around 100,000 and probably higher, not between 8,000 and l94,000 as Mr. Straw states. While far higher than the Iraq Ministry of health surveillance estimates, on 17 August the minister himself described surveillance in Iraq as geographically incomplete, insensitive and missing most health events.We, the occupying nations, should aspire to acknowledge the dignity of every life lost, and to monitor trends and causes of deaths to better serve the Iraqis, and in doing so, sooner end this deadly occupation.Les Roberts, Gilbert Burnham Centre for International Emergency, Disaster and Refugee Studies, Johns Hopkins Bloomberg School of Public Health, Baltimore; Richard Garfield School of Nursing, Columbia University, New York, USA

Mike;

You are right that originally didn't say: "there was " no way " the war was responsible for the increase", I was paraphrasing, you said:

"It's difficult to make the argument that the 7 men and children who died from heart attack and stroke can be attributed to invasion, when we have 3 women dying before invasion of the same cause, and only 1 after."

Your point was that the increase in heart attacks was a cause of death that didn't count. The blame couldn't be pinned on the invasion, so the total estimated excess mortality figure for increased deaths after the invasion had to be revised downwards. I think I've cast serious doubt upon this.

"The data shows there was an increase in heart attack and stroke fatalities among the 7,800 people that made up the sample, but you're trying to pass this off as an accurate extrapolation for the entire country. In effect, you're saying, " I like the numbers from this particular sample, they assist my argument, so therefore they're proof that the ratio holds true for all of Iraq." The problem is, they can't be representative for purposes of extrapolation."

I wasn't attempting to make any extrapolation across all of Iraq. You had commented that you thought it could have been chance that heart attacks in the sample increased. I was saying that it is unlikely that that heart attack deaths (in the sample) after the war could have been generated by the risk given by the data before the war. It is unlikely to have been chance that caused the increase.

I'm not going to do the maths to extrapolate from the data to the population of Iraq. I don't have all the data for this anyway. But based upon what data we do have I can certainly say it is more likely that heart attacks mortality increased after the War, rather than falling or staying the same.

"Finally, we have women dying at a rate of 3 to 1 BEFORE regime change... It makes no sense, and conflicts with your assumption that fatal heart attacks have risen dramatically, unless we accept the premise you put forward earlier, that wartime stress affects children and men more than women and the elderly. Are you sure you want to go there?"

The data showing that heart attacks for women in the sample fell, does not conflict with my "assumption" that fatal heart attacks have rose dramatically. It is possible that overall heart attack deaths rose but heart attacks among women didn't. It also isn't an *assumption* that fatal heart attacks rose dramatically, the sample data shows they did.

I'm quite happy with the idea that stress converted to heart attacks among women differently from men. I've already explained that the elderly data point is consistant with that for children and men. I don't think this is a problem for me.

"I'm not discounting there could be an increase in heart attack mortality, and that some of it, possibly even all of it, could be caused by the war. But the study shows a 61% increase in heart attack mortality for the entire sampling. We have no idea, from this study alone, if this is accurate for the entire country... Additional studies would provide a clearer picture, and the actual increase, if any, could be much lower."

Of course - based on the low data - we're not going to be sure about the exact change in heart attack deaths in Iraq. I'd welcome further studies too. The point of the Lancet study wasn't to collect data on heart attacks (and I'm sure that'd be impossible), it was to estimate excess mortality after the Invasion. I've criticised the study above. But I think the extrapolation technique is valid, and don't think that the rise in heart attack deaths should somehow "not count" insofar as it contributes to the estimated excess mortality.

This letter from the authors of the study defending their findings was published December 12 in the UK newspaper The Independent on Sunday:Iraq toll: at least 100,000 You reported ("Polish hostage held in iraq is released unharmed", 21 November) the Foreign Secretary's response to our study published in The Lancet of civilian deaths in Iraq. It is heartening that Jack Straw has addressed the topic in such detail. However, his response includes an apparent misreading of our results.Our study found that violence was widespread and up 58-fold after the invasion; that from 32 of the neighbourhoods we visited we estimated 98,000 excess deaths; and that from the sample of the most war-torn communities represented by 30 households in Fallujah more people had probably died than in all of the rest of the country combined.Fallujah is the only insight into those cities experiencing extreme violence (ie Ramadi, Tallafar, Fallujah, Najaf); all the others were passed over in our sample by random chance. If the Fallujah duster is representative, there were about 200,000 excess deaths above the 98,000.Perhaps Fallujah is so unique that it represents only Fallujah, implying that it represents only 50-70,000 additional deaths. There is a tiny chance that the neighborhood we visited in Fallujah was worse than the average experience, and only corresponds with a couple of tens of thousands of deaths. We also explain why, given study limitations, our estimate is likely to be low. Therefore, when taken in total, we concluded that the civilian death toll was at least around 100,000 and probably higher, not between 8,000 and l94,000 as Mr. Straw states. While far higher than the Iraq Ministry of health surveillance estimates, on 17 August the minister himself described surveillance in Iraq as geographically incomplete, insensitive and missing most health events.We, the occupying nations, should aspire to acknowledge the dignity of every life lost, and to monitor trends and causes of deaths to better serve the Iraqis, and in doing so, sooner end this deadly occupation.Les Roberts, Gilbert Burnham Centre for International Emergency, Disaster and Refugee Studies, Johns Hopkins Bloomberg School of Public Health, Baltimore; Richard Garfield School of Nursing, Columbia University, New York, USA

James Brown writes, "Mark Bahner gets it right !"

James, that's not something that needs to be bolded and have an exclamation point. "Mark Bahner gets it wrong!" would be something that needs an exclamation point. And a question mark. ;-)

James Brown continues, "Anyone who disagrees= troll... from there, it is perfectly easy to understand why Tim doesn't see any evidence that contradicts his views. No evidence has been provided that undermines the Lancet study- except from trolls, and trolls don't count."

Yes, but it's even *worse* than that, from the standpoint of Tim Lambert's rationality/credibility/honesty. Tim Lambert doesn't just ignore evidence that undermines the Lancet study. He even makes totally bogus claims that NO rational/credible/honest reader could make, based on the Lancet study. For example, his claim that 95%+ of the "excess" deaths outside Fallujah were civilians.

That is simply pure BS.

"Tim even gets to make stuff up !"

Yes, exactly. He makes up pure BS. And then he claims to be different from other people who he says are making up pure BS. The irony of his post titles like, "La la la I can't hear you" is incredible. (Kind of funny, too.)

Jeff Harvey wrote: "Many of the contributors to this blog continue in the rich tradition of western denial and complicity."
jeff, I think you have two real problems in your argument.
1. That everyone who disagrees with the lancet study supports the US war in Iraq. This is wrong.
2. The science must stand on its own two feet; it must be able to support the claims that it makes. Yet we know that even on its own terms, the authors themselves do not believe that their sampling methodology was sufficiently robust to be accurate for extrapolating to the whole population. They excluded the Falluja cluster !
they do not trust their own study- so why should you ?
j

Mark Bahner wrote:"The irony of his post titles like, "La la la I can't hear you" is incredible."
Hi mark
I have to say I was amused by his title, and also by his new principle of statistics...
Tribbs' post shows the authors of the study picking and choosing which parts of the study they want to accept, and disavowing the statistical confidence limits of their own study. It's hilarious !
j

Enquiring mind:

"I'm quite happy with the idea that stress converted to heart attacks among women differently from men. I've already explained that the elderly data point is consistant with that for children and men. I don't think this is a problem for me."

On what evidence are you trying to justify the notion that women somehow handled the stress of war better than men and children, in relation to heart attack mortality? You're doing so only because a sampling of 7,800 people has provided this quirk, and now you're grasping at straws to try and explain it away.

And no, the elderly data is not consistent with the men and children. The elderly went from 8 to 10, the other 2 groups from zero to 5 and zero to 2 respectively. they aren't comparable. The small difference between the elderly rate could be entirely accounted for by the circumstances I raised before, the fact that the elderly are much more likely to die of heart attack, they have a 3 month longer period after the war to die from it, and there is a slight increase in their numbers after regime change. You have no grounds for claiming a jump of 8 to 10 in this demographic is the result of war-induced stress. It isn't there.

"You had commented that you thought it could have been chance that heart attacks in the sample increased. I was saying that it is unlikely that that heart attack deaths (in the sample) after the war could have been generated by the risk given by the data before the war. It is unlikely to have been chance that caused the increase."

I believe both the increase and/or the rate of increase could be chance (and therefore not representative for extrapolation), and I also believe, as I've said several times, that some or all of the actual deaths contained in the study may not be attributable to the war.

When discussing this, you have to confine the debate to the actual 18 deaths in the study. Simply put, each of these 18 deaths has to be evaluated based on one criteria; would the deceased still be alive if the coalition didn't invade? We have no way of determining this from the study data. Another study, or studies, might give us a clearer picture, but this one can't, on its own, because of its demographic breakdown.

You say you weren't trying to make an extrapolation from the excess heart attack deaths, but the study does, and it has a bearing on the overall excess death number. There really is no point in arguing this with me if you aren't trying to infer an extrapolation. There may well be 3 or 4 heart attacks from this sampling that wouldn't have occurred without regime change. There may be more, there may have been none. The study doesn't tell us, but it leaves us with a 61% increase in heart attack deaths, which appear to be lumped in with all other causes of excess death and dumped on the doorstep of the coalition.

I have to ask myself, if Saddam and his regime had been overthrown by the Iraqi people in a bloody civil war, would anyone have bothered to conduct a mortality study that included the use of heart attack and accident deaths to pile up a number designed to shame the overthrowing populace?

Jeff:

I've read through all of the Lancet threads at this site, and all seem to contain at least one comment from you that sounds like a broken record. Your terminology doesn't even change; "denialists like alcoholics, unpeople. " I'm not interested in talking about the Americans in the Phillipines a hundred years ago, or how much you despise the U.S. in general, when this thread concerns the Lancet study.

You thank Tim for his work in defending the Lancet study, yet repeat the propaganda that more than a million Iraqis were killed by the coalition during and after the First Gulf War, and that the country was flattened. You fail to realize that most of the knowledgable defenders of the Lancet study reject the scenario you paint, that Iraq and its infrastructure had been bombed to smithereens. It doesn't fit with the pre-war analysis by the authors of the situation in Iraq, nor would it allow a defence of the study, as the coalition could hardly have made things worse if we adopt your script.

Concerning the link you posted, this " reporter, " Dahr Jamail, is not an objective commentator in the least. One site (a favourable one to Jamail) called him a "writer and political activist." I think that's being kind, but it's closer to the truth than " reporter."

Here's an example of Jamail's impartiality:

CS: "The US Corporate media consistently characterizes the Iraqi resistance as "foreign terrorists and former Ba'athist insurgents". In your experience, is this an accurate portrayal? If not, why?"

DJ: "This is propaganda of the worst kind. Most Iraqis refer to the Iraqi Resistance as "patriots." Which of course most of them are-they are, especially in Fallujah, primarily composed of people who simply are resisting the occupation of their country by a foreign power. They are people who have had family members killed, detained, tortured and humiliated by the illegal occupiers of their shattered country."

I spent some considerable time today reading about this guy on the net. I can see why you like him so much Jeff. I can also see why he's claiming that:

" Nearly every refugee from Fallujah I've interviewed has spoken of mass executions, tanks rolling over the wounded in the streets, bodies being thrown in the Euphrates by the military, and other atrocities." (He's referring to the U.S. Army doing the mass executions, by the way)

I've also found references to Jamail accusing the U.S. of using poison gas in Falluja, secretly burying their own dead (U.S. soldiers) in mass graves in Iraq to hide their true casulty numbers, and referring to the 4 dead contractors in Falluja as " militants," while at the same time acknowledging that they were delivering food.

You can treat Jamail's rants as gospel, I'll stay in "denial," thanks.

If you're interested in a full on debate over Iraq and Bush-Hitler, I'm game, but not here and now. I've been here for a week and a half debating the Lancet study, and I'd prefer to finish it first.

Ian Gould:

In relation to your comment that the Iraqi medical infrastructure had " collapsed," and therefore an increase in heart attack mortality was inevitable, I tried to get this interview linked, but the CBC Radio homepage advises that all its archived clips past April 2004 are no longer available online. The interview was a 6 minute clip with Doctor Morten Rostrup, the head of Medecins sans Frontiere, and was given on April 25th, 2003, shortly after regime change. Here's how the CBC heads it up at their website:

"Doctors without borders has decided that there is little need for their services in Iraq. Morten Rostrup is the international president of Doctors Without Borders. He was in Baghdad throughout the war. Today, he's back in Oslo."

I had transcribed a few quotes from the interview, but can't find them now. The gist of the interview was that while the Iraqi medical system had some difficulties, it wasn't in a shambles after regime change, there was no humanitarian crisis, and therefore Medecins sans frontiere wasn't going to stay in Iraq. Rostrup accused some who opposed the war (unnamed) of trying to create the impression that there was a health care/humanitarian disaster in progress. Coming from the head of a left leaning NGO that isn't known for saying nice things about the U.S., I was startled by his comments.

This was the situation Rostrup noted just after the fall of Saddam. I won't argue that in some parts of Iraq since then, at various times, the health care system in Iraq has been degraded by the instability in certain regions. I just find the generalization that the health care system has " collapsed," to be inaccurate.

Donald:

I don't see much point in responding to what you've posted, in light of the letter Tribbs posted, which clarifies the study authors' opinions concerning the extent of deaths in Fallujah. It goes without saying that I'm more than a little skeptical of their view. In any event, I've posted something about this at the newest thread, if you wanted to come over there to talk further.

Thanks

Let's look at soem of the other things rostrup has said
abotu Iraq.

April 2003:

In addition to this, we also know there are many patients there with chronic diseases. Diabetes, epilepsy, cardio-vascular disease, kidney diseases that need regular follow up and also medication. Many of these patients have been without their regular medication and if you think of a diabetic who is insulin dependent, without insulin you will die of diabetes if you do not get care.

So what we still do not know is what has been going on in the many homes in the Baghdad, a city with millions of people, when the primary health system along with the hospitals broke down. But I hope we shall be able to do a proper assessment and the teams are trying to do this as well, even though it is a bit difficult under the current situation in Baghdad.

http://www.doctorswithoutborders.org/news/2003/morten_press_4-23-2003.s…

May 2003

<this lack of protection led to a total disarray later on in the hospital situation. And that I think is partly the cause of the very difficult situation we even see today.>> Emphasis added

"more than three weeks after the fall of Baghdad we still see most of the hospitals in a total disarray. We still see that they are disorganized. We still see that emergency rooms are filled with patients. That operation theaters are not fully functioning. In fact there is not one hospital in Baghdad today fully operational."

http://www.doctorswithoutborders.org/publications/other/iraq_pressconfe…

Here's what MSF Director Gerik Ooms said about Iraq in November 2004:

"It has become impossible for MSF as an organization to guarantee an acceptable level of security for our staff, be they foreign or Iraqi.We deeply regret the fact that we will no longer be able to provide much needed medical help to the Iraqi people."
http://www.doctorswithoutborders.org/pr/2004/11-04-2004.shtml

I beleive that you are quoting a single soruce that is over 18 months old which contradicts numerous later statements from both the same source and others.

Mike;

Sorry to keep banging on about this.

On what evidence are you trying to justify the notion that women somehow handled the stress of war better than men and children, in relation to heart attack mortality?

I'm not trying to justify the notion that women somehow handled the stress of war better than men. It's not in the slightest bit relevant to the point I'm making, which is that it's totally legitimate to attribute excess cardiac mortality to the War. It's actually just a distraction to it.

And no, the elderly data is not consistent with the men and children. The elderly went from 8 to 10, the other 2 groups from zero to 5 and zero to 2 respectively. they aren't comparable... You have no grounds for claiming a jump of 8 to 10 in this demographic is the result of war-induced stress.

I never claimed the rise in deaths in this group could be attributed to stress. You asked why there wasn't a "dramatic increase in deaths from heart attack among the elderly", if stress could cause adult men to die from cardiac death. You clearly haven't engaged with the point I made above. Let me spell it out for you.

Using rough figures. The chance of dying for men increased after the war by 1 in 550. The chance for dying for the elderly, before the war, was 1 in 33. That's one reason why there wasn't a dramatic increase, the elderly were already dying at a very high rate before the war. An increase in the elderly mortality rate the same as the increase for the men wouldn't be dramatic, and given that there are only 331 elderly, it also wouldn't be detectable.

When discussing this, you have to confine the debate to the actual 18 deaths in the study. Simply put, each of these 18 deaths has to be evaluated based on one criteria; would the deceased still be alive if the coalition didn't invade?

This is the fundamental difference between us. I don't think you do need a case by case accounting. This just degenerates into a series of hypothetical questions, and is impossible to answer in theory, let alone in practice. Given that major disruption has been caused by the war, comparing deaths before and after the invasion is perfectly legitimate. Excess deaths can be attributed to the war.

You say you weren't trying to make an extrapolation from the excess heart attack deaths, but the study does, and it has a bearing on the overall excess death number... The study doesn't tell us, but it leaves us with a 61% increase in heart attack deaths, which appear to be lumped in with all other causes of excess death and dumped on the doorstep of the coalition.

The study doesn't just extrapolate from heart attacks. It extrapolates from a random (and much larger) sample of all deaths before the war to deaths in that same sample afterwards. Heart attack deaths are included, I don't think you've given any convincing reason they shouldn't be.

Enquiring mind:

You wrote:

"I never claimed the rise in deaths in this group (the elderly) could be attributed to stress."

But you did, in an earlier post, with this statement:

"I'm quite happy with the idea that stress converted to heart attacks among women differently from men. I've already explained that the elderly data point is consistant with that for children and men."

Then, you said this in your latest post:

This was from an earlier post:

"I'm quite happy with the idea that stress converted to heart attacks among women differently from men."

And this one:

"It's entirely possible that deaths rose among men and children because these two groups had stress elevated most by the war,....."

You're contradicting yourself on this subject.

Enquiring mind, you keep trying to claim that the study data is all the proof required that heart attack mortality increased dramatically after the war, and further that the increase is directly attributable to the war. I've given you my reasons for believing the study data isn't capable of providing this proof. I'm not going to bother repeating my arguments. You've dug in your heels on the subject, and nothing I can say is going to sway you. That's fine, no hard feelings. I'm not going to debate it any further, time is precious, and there's no point to continuing. I'll leave the last word to you.......

Enquiring mind:

My last post omitted a quote from you, and therefore won't make sense. When I stated "Then, you said this in your latest post:", I neglected to include the relevant quote from you. Here it is:

"I'm not trying to justify the notion that women somehow handled the stress of war better than men. "

Ian:

I hope you're not suggesting that I'm playing fast and loose with the truth here, in relation to what Rostrup said in the CBC interview. Here's a quote from him, from the first link you posted:

"But I think it is telling that, after two weeks, and after being in ten cities, MSF has not found any reason to justify a major humanitarian medical program in Iraq. I think it is pretty interesting and for me what we have been doing so far contrasts what has been portrayed, especially in Europe I think, about this huge humanitarian crisis or catastrophe in Iraq. This perception does not correspond with what we have been so far able to see ourselves in Iraq and I think we also have to ask a question whether it is really justified to use the whole notion of a major humanitarian catastrophe from a medical perspective at this point in time."

Isn't that pretty much exactly what I paraphrased Rostrup had said in the CBC interview?

Here's what Rostrup said in the second link you provide, that counters your original suggestion that there was a " collapse of the medical infrastructure.":

"But it seems pretty clear to me that there is enough staff in Baghdad today, enough doctors and nurses, also medical supplies seem to be okay for the time being, there are some specific needs in the early phase but still it is not very bad and what is lacking is really organization of the hospitals and the immediate care of the patients."

3 weeks after the invasion, Rostrup said that none of the hospitals in Iraq was " fully operational." How do you go from there, to the conclusion that the medical infrastructure had collapsed? "But still it is not very bad....." is a plain summary of the medical situation in Iraq as Rostrup saw it. If you take the totality of Rostrup's statements from both links you provide, your statement is false.

Your third link in no way suggests a collapse of medical infrastructure in Iraq. It deals with the withdrawal of MSF from Iraq last month due to safety concerns for its staff. I don't blame them for leaving. With the murder of Margaret Hassan, it proves that the Iraqi " Minute-men " will not distinguish between westerners providing completely selfless aid to the Iraqi people, and those westerners who are combatants. MSF left because the staff were possible targets.

I find it curious that you would dismiss my reference to Rostrup's comments because " it is over 18 months old," yet the only 2 cites you mention that are relevant are from exactly the same time period.

I mentioned Rostrup's CBC interview because it refuted your claim that Iraq's medical infrastructure had collapsed. You countered with two other interviews where Rostrup denies that Iraq's medical infrastructure had collapsed. You should avoid gratuitous generalizations that fail to withstand scrutiny.

What is your definition of "collapsed"?

Fairly obivously it doesn't incldue a situatio nwhere no hospitals are functioning normally and where health workers are afraid to do their jobs.

If there weren't extrmeely serious problems with the Iraqi health system why were MSF still working there up to November, 2004>

So what would you require to justify that a statement that the medical system has collapsed: every single hospital in the country leveled to the ground and every doctor dead?

If that did happen you probably counter: "But there are still band-aids for sale in some shops therefore the health system hasn't collapsed."

With a budget of twenty million Dollars, the Iraq health system was comatose before the intervention. They now get fifty times the budget that was available under Saddam.

For me the discussion on the Lancet study is pretty finished, unless I hear some good argument to contradict my view that the study adds nothing new and effectively is propaganda that serves the interests of fascist beheaders and criminals.

Let me add something on the bombing of Dueren compared to that of Fallujah. Pictures don't describe well enough what happened to my home town in WWII. 150,000 bombs were dropped in the space of 36 minutes, 8 per remaining inhabitant. Most of the bombs were firebombs and the place went up in a firestorm. People in the city centre who survived did so, because they had gas masks and underground tunnels available to get from one basement to the next. Those who stayed in the basements under the fire storm died eventually through asphyxiation.

Fallujah is nothing like that, I don't believe for a minute that tens of thousands of civilians have died from coalition bombing in that city. I believe it's at most a few thousand, and may be as low as a few hundred, and I lay the entire blame for that on the "insurgents".

Which is, interestingly enough, something where I sort of disagree with Mike. I haven't seen reliable evidence that would indicate that coalition bombing has been disproportionate, or for that matter an important cause of civilian death.

As I've said before, my absolute main concern is that the troops are not coached sufficiently to be culturally aware, that includes learning Arabic well, and extensive contact with Iraqis in safe surroundings.

Mike;

You list four comments by me:

(1) "I'm quite happy with the idea that stress converted to heart attacks among women differently from men. I've already explained that the elderly data point is consistant with that for children and men."

(2) "I'm not trying to justify the notion that women somehow handled the stress of war better than men."

(3) "I'm quite happy with the idea that stress converted to heart attacks among women differently from men."

(4) "It's entirely possible that deaths rose among men and children because these two groups had stress elevated most by the war,....."

And say they are instances of me claiming that the (increase of 2) deaths in the elderly could be attributed to stress. They don't. I may be being dense here; so point it out if I'm missing something obvious.

My point was that it's totally legitimate to attribute totaly excess cardiac mortality to the war. Your point was that the idea that war could raise mortality from heart attacks was not credible, and deaths from this cause should be discounted for the purpose of estimates. So the study's totals should be revised downwards. You are wrong. You've even stopped trying to defend this position some time ago.

"you keep trying to claim that the study data is all the proof required that heart attack mortality increased dramatically after the war, and further that the increase is directly attributable to the war."

I'm not claiming that the study proves heart attack deaths rose in Iraq - I don't have the data to make this inference. I am claiming that heart attack deaths rose in our data. I am claiming that this increase is unlikely to be from chance. I am claiming that this should be included in the data used to generate the total post-War mortality estimates (unlike your earlier claim). I am also claiming that it is a well established finding that war stresses people which increases heart attack mortality. If you were to place a bet on what happened to heart attacks in Iraq post-War, you'd bet it went up.

Being tortured in Saddam's prisons is also likely to lead to deaths from heart attack. The study didn't even sample the prison population.

The numbers for mortality ex violence in the Lancet study have a basis that is far too weak to conclude very much from them at all (overall the evidence, including information apart from the Lancet study, indicates, in my opinion, that so far there has been little change in mortality ex violence, with nearly equal likelihood of decline as increase). They are also of a completely different nature. Half of all smokers die eventually from smoking. But, the kind of mass dying this implies is rather different from the Nazi death camps, or soldiers dying in war.

Aggregating murders and other kinds of death is a tricky business, and it's something one ought to be particularly careful of, when attributing moral responsibility.

Furthermore, excess deaths can disappear, war dead cannot. If infant mortality gets measured every year for the next 10 years, we might find half a million children alive that wouldn't have been at the infant mortality prevalent before the war. There a 100,000 dead (as in excess mortality) can quickly turn into a 100,000 saved. Within the space of a year or two even.

That's not what people mean when they refer to war dead. What they are talking about is those directly killed by bombs and guns.

Heiko:

When you say:

" Which is, interestingly enough, something where I sort of disagree with Mike. I haven't seen reliable evidence that would indicate that coalition bombing has been disproportionate, or for that matter an important cause of civilian death."

I'm wondering what the context is for the disagreement? I'm not being snarky, I just wanted to know the terms of reference.

I made a point in the most recent thread, to Donald Johnson, that I am very critical of the utilization of coalition bombing, particularly the April bombing of Falluja. I don't believe that several tens of thousands have died from coalition bombs, but 10,000 wouldn't surprise me in the least, and the number could be significantly higher. I think what you and I are both in agreement is the opinion that the authors' conclusion that most of the violent excess deaths were caused by bombing can not be corroborated by their own data.

Ian:

"So what would you require to justify that a statement that the medical system has collapsed: every single hospital in the country leveled to the ground and every doctor dead?"

Yes, that would constitute a collapse of the medical system. So would a scenario far less catastrophic. That hasn't happened, nor has anything remotely close to it happened. If you have some evidence that it has, then let's see it. You have made the allegation that the medical system has collapsed, but have offered no proof of any kind.

The majority of Iraq is experiencing virtually no violence from the insurgency, or only sporadic spasms of violence from isolated car bombings. The insurgency is centered in Anbar province, with sporadic, sometimes sustained, violence in Baghdad as well. There is nothing to cause the collapse of the medical system outside these areas. That's why it hasn't collapsed, along with Heiko's observation that hospitals have far greater funding and avenues to obtain funding than the Saddam years.

The time frame for greatest peril to the Iraqi health system was during the invasion and immediately after. Rostrup's words for that time period speak for themselves. How do you rationalize a " collapse " of the health care system, in light of the quotes I've provided from Rostrup, from your very own links!

As I said earlier, you decided to make a gratuitous generalization, and you were called on it.

I rationalise "collapse" fromn Rostrup description of the system as in "total disarry".

So let's move beyond the semantic games you want to play. Let's call the Iraqi hopsital system "slightly less than ideal' (SLTI) and afdmit that maybe just possibly bombing hopsitals and the power and water plants that service them, allowing rampaging mobs to strip them of everythign to value and swamping them with thosuands of additional serious cases might actually somehow make them more SLTI.

Let's see if there's evidence of this SLTI status:

http://www.usatoday.com/news/health/2004-08-18-iraq-hospitals-usat_x.htm "When Iraqi physician Mahmud Thamer stepped down from the U.S. military helicopter that carried him into Baghdad on June 6, 2003, after 34 years of exile, he found Iraq's health system in shambles" ...
"The day after the two men arrived in Baghdad's blast-furnace heat, they found the health ministry all but gutted."...Emergency services were overwhelmed. "This is especially true because you have attacks all over the country," says Thamer, now of Johns Hopkins University. "They never anticipated these kinds of patient loads that would break any hospital's back."

http://www.occupationwatch.org/article.php?id=3050At Baghdad's Central Teaching Hospital for Children, gallons of raw sewage wash across the floors. The drinking water is contaminated. According to doctors, 80 percent of patients leave with infections they did not have when they arrived.

So the STLI Iraqi system hasn't "collapsed" it's just "in chaos" and is being run by a ministry that's been "gutted"

Now let's leave the semantics behind, the STLI Iraqi hospital system in the months following the invasion (sorry would you prefer "glorious and wholely justified liberation by their peace-loving american benefactors"?)would bel ikely to result in an increased number of deaths from heart attacks even if the incidence of heart disease didn't increase (unless you're going to argue that hospitals work more effectively when awash with raw sewage, without electric light because all the cabling has been ripped out and with armed gangs invading the hospitals on a daily basis).

Men are more prone to heart disease than women, therefore a higher increase in the male mortality rate from this cause than the female is likely to be expected.

Here's how Amnesty International described the medical situation in Iraq immediately after the invasion:

http://web.amnesty.org/library/index/engmde140852003Beginning in Basra on 7 April, followed by Baghdad on 9 April and Kirkuk the following day, crowds of desperate people have taken to the streets, looting, burning and destroying government offices and, more ominously, institutions vital to their future, including schools, universities and hospitals. ...

The International Committee of the Red Cross (ICRC) said it was concerned that the looting was further hampering attempts to deliver humanitarian assistance. UNICEF's representative to Iraq, Carel de Rooey, expressed similar fears when he told a press conference on 9 April: "What is terribly worrying about the looting, chaos and breakdown of order is that the systems we counted on may completely collapse."

On 6 April UN relief agencies warned that a health crisis faced the five million inhabitants of Baghdad, with hospitals overwhelmed and infrastructure devastated.

The same day the ICRC reported that emergency services were not being provided by the hospital in Umm Qasr and patients could not travel safely to Basra for treatment.

By 9 April the ICRC was describing the humanitarian situation in Baghdad as "critical", particularly as health, power and water workers could not reach their workplaces safely, and many medical centres lacked water and electricity.

On 10 April the ICRC resumed work in the capital but reported that al-Kindi hospital in Baghdad had been ransacked and that street violence and looting had forced the closure of others.
The inability of ambulances and other vehicles to move freely in the capital was highlighted on 10 April by an AFP photographer, who reported that around 20 bodies, including those of children, were still strewn on the road between al-Dora and the airport, days after they were killed.

As to the increase in the Iraqi medical budget:

1. I am referring to the period during and immediately after the invasion, subsequent increases in the budget even if they were having a positive impact on the ground are irrelevant to the death rate during that period

2. Only a small part of the budgeted spending on reconstruction has actually occurred.

3. much of the increase in the medical budget is attributable to the icnrease in salaries. most of this increase actually reflects the change from the previous system where government workers received free accommodation, food and power. Requiring these workers to pay for these services and increasing their salaries to cover their higher expenses odesn;t increase medical resources.

Okay, let's see if I can avoid this response being eaten like the last one. and if the formattign survives the cut-and-paste from Word.

Mike wrote: "The majority of Iraq is experiencing virtually no violence from the insurgency, or only sporadic spasms of violence from isolated car bombings. The insurgency is centered in Anbar province, with sporadic, sometimes sustained, violence in Baghdad as well. There is nothing to cause the collapse of the medical system outside these areas."
This statement conflates several distinct factual and logical errors.

1.The insurgency isn't the only possible cause for the collapse - sorry, for the SLTI status of the Iraqi medical system. The principal cause seems to have been the widespread looting and wave of violent crime which swept across the country in the immediate aftermath of the fall of Saddam. Other likely contributing factors include the pre-war dilapidation of the medical system, bombing and other war-related damage to hospitals, the destruction of large parts of the electricity and water treatment systems and the sacking of hundreds of doctors and Health Ministry employees for Baathist affiliations.
2.This discussion was focused initially on the Lancet Report which is historic look at excess deaths over the whole post-invasion period. There have been periods of intense fighting right across Iraq including the extremely widespread fighting in May and the subsequent fighting in Najaf and Karbala. Even if your attempt to characterise the insurgency as currently, largely confined to Anbar were correct, and I will show it isn't, it's misleading in this context.
3.There may be 19 provinces in Iraq but Baghdad and Anbar make up around 1/3 of the country's population. Your argument here is roughly equivalent to saying "The fighting in Australia is centred in Sydney with occasional outbursts of fighting in Melbourne, much of the rest of the country is largely peaceful." Additionally, as in every country in the world, medical services in Iraq aren't evenly distributed around the whole country, they're concentrated in larger population centres, like Baghdad. This is even before we consider the likelihood that Saddam would have put a disproportionate amount of medical resources into Sunni-majority areas such as Baghdad and Anbar province.
4.Is the majority of the violence confined to Anbar province? I can't find a source listing all violent incidents in Iraq but I can find to useful proxies: the list of American military casualties at Antiwar.com (http://antiwar.com/casualties/list.php) and the list of Iraqi civilian casualties on the Iraqi Bodycount website http://www.iraqbodycount.net/database/bodycount_date_down.php. In both cases I'm going to look at data for the month of December to date. If you want to argue that data from a longer period should be examined, you're quite welcome to do so.
5.38 American personnel were killed in Iraq in the period 1-15 December (or more accurately 38 deaths were reported for that period. Some of those deaths may be related to injuries received earlier.). 5 of those deaths were non-combat-related. Of the 33 combat-related fatalities, only 18 or just over half occurred in Anbar province. (I was unable locate either Habbaniyah or "Camp Cedar" so I have included these figures in the Anbar figure to err on the side of caution.) A further 6 combat-related deaths occurred in Baghdad, leaving 9 deaths or almost a third occurred in the rest of Iraq which you claim is experiencing "virtually no violence".
6.Iraq Bound count records 78 civilian deaths in the first ten days of December - which for reasons discussed elsewhere on this blog is almost definitely a significant underestimate. Only 8 of the casualties recorded occurred in Anbar province. Of the remaining 70, 48 occurred in Baghdad and 22 occurred elsewhere.
Iraqi Body Count depends largely on English language press coverage for its data. given the concentration of western media in Baghdad, these figures are liekly to underreprot violence elsewehere in Iraq.
Are these figures 100% accurate? No.
If you can find better data feel free to quote it.

Ian:

Here's that quote again from your own first link from Rostrup:

"I think it is pretty interesting and for me what we have been doing so far contrasts what has been portrayed, especially in Europe I think, about this huge humanitarian crisis or catastrophe in Iraq. This perception does not correspond with what we have been so far able to see ourselves in Iraq and I think we also have to ask a question whether it is really justified to use the whole notion of a major humanitarian catastrophe from a medical perspective at this point in time."

As you'll recall, I posted this quote earlier. What I want to draw your attention to is the criticism Rostrup reserves for those who are attempting to portray the medical situation in Iraq as a catastrophe:

"I think it is pretty interesting and for me what we have been doing so far contrasts what has been portrayed, especially in Europe I think, about this huge humanitarian crisis or catastrophe in Iraq."

Ian, Rostrup is talking about you here as well. In trying to use Rostrup's words to bolster your case, you're actually engaging in the same exaggeration of the situation that Rostrup objects to. Frankly, I can't come up with anything more damning to your assertion than what you yourself have provided. If you were willing to debate this question fairly and honestly, this would end the debate. I don't suspect that is the case, so I'll address some of your points.

You can't rationalize " collapse " from Rostrup using the expression "total disarray," to describe the days immediately following regime change. Even more important, you cannot ignore (although you're doing your best) the quotes I've found from Rostrup that utterly refute your claim of a collapse of the medical system. The Lancet study measured deaths for a 17 month period following invasion. You're trying to pass off the immediate chaotic aftermath of regime change as being representative of the entire post-war period covered by the study. That simply isn't true.

"So let's move beyond the semantic games you want to play."

Ian, please. Spare me your feigned righteous indignation. You were the one playing the semantics game. There is a huge difference between a claim of "collapse" of the Iraqi medical system, and the actual current state of this infrastructure. That isn't a matter of semantics, however much you'd like it to be.

The Lancet study itself directly contradicts your claim of a " collapse.":

"It is surprising that beyond the elevation in infant mortality and the rate of violent death, mortality in Iraq seems otherwise to be similar to the period preceding the invasion. This similarity could be a reflection of the skill and function of the Iraqi health system or the capacity of the population to adapt to conditions of insecurity."

A collapse of the medical infrastructure in Iraq inevitably and unquestionably would result in a dramatic rise in the general mortality rate (excluding violence) that the authors are referring to. The authors are telling you Ian, in plain language, that this has not occurred.

Isn't this the argument that was used prior to the war by those who opposed sanctions? That the huge numbers of excess deaths purported to have occurred in the sanctions era were largely the result of a non-functioning, sanctions-devastated medical system? You can't have one without the other, and the authors are telling us they don't see evidence of either, let alone one.

With regard to the links you've provided. The first one, the USA Today story mentioning Mahmud Thamer, reveals you playing the same game you did with the Rostrup links, omitting much that contradicts your argument. Why do you do this?

You chose to highlight from this article as your main point the fact that the Iraqi Health Ministry headquarters had been " gutted " during the fighting. Later, returning to this same point, you mock:

" So the STLI Iraqi system hasn't "collapsed" it's just "in chaos" and is being run by a ministry that's been "gutted"

The only problem is, you left out this quote from the article:

"The building had to be completely renovated, a task that took until late fall." (2003, that is).

So no Ian, the Iraqi system isn't being run from a ministry that's been "gutted."

This article portrays the Saddam years as primarily responsible for the problems inherent in the Iraqi medical system, and as the article progresses, it notes an improving situation after regime change, not a deteriorating one. Here are some passages from the article that demonstrate this:

"The decline that began two decades ago accelerated in the '80s, when Saddam diverted the health budget to pay for his wars with Iran and Kuwait. The trade embargo of the '90s took a toll, too, though there's evidence suggesting that Saddam drained resources from maternal and infant health to boost the number of deaths "so he could blame the United Nations' sanctions," Haveman says.

In 2002, the Iraqi government health budget was $16 million, or 64 cents per person. By January, the coalition boosted that amount to $40 per person, using $1 billion in oil revenues. That doesn't include $793 million devoted to health in President Bush's $18 billion supplemental request approved and signed last fall."

"Last week, in a meeting sponsored by the U.S. Department of Defense and the quasi-public Institute of Medicine of the National Academy of Sciences in Washington, D.C., officials offered their assessments of how the process is proceeding.

"It's been a very short time," says former minister Abbas. "At least we kicked the ball and got it rolling."

Your other 2 links don't work. The 2nd one brings me to " Iraq Occupation Watch," with numerous links, but I have no idea where your article is. From the searching I did do at this site, it's apparent this is a virulently anti-American propaganda entity, with absolutely no interest in objectivity or fairness. Your 3rd link is a dead link. In any case, the text you provide from the 3rd link only covers a 3 day period, right at the time of the capture of Baghdad!

While I think your attempt to prove the insurgency isn't largely confined to Anbar and Baghdad is getting ridiculously far away from the debate, I did dig up some stats independent of yours, from the U.S. Government's Department of Defence Homepage. Their press releases identify the circumstances and locations of each soldier killed in Iraq. My count for deaths in combat in December is slightly different than yours. I actually get 2 more combat deaths than you (35). Of these, at least 20 are in Anbar Province (13 listed as Anbar, 3 in Ramadi, 3 in Habbaniyah, and 1 in Khalidiyah, all of which are definitely in Anbar). Your " Camp Cedar " death is in a place called Taji, which as near as I can place it on the governorate maps, is either just inside the Anbar boundary or just outside it, too close to divorce it from Anbar, in my opinion. I count 5 dead in Baghdad, and 3 in Babil province, which is one of the smallest governorates, bordering on both Anbar and Baghdad, and incidently borders on Anbar close to Falluja and Ramadi. Baghdad and these 2 provinces account for 29 of the 35 deaths from the research I've done. The remaining deaths were 3 in Mosul, 2 in Diyala province, and 1 in Kirkuk.

Needless to say, my figures tell a different story than yours.

For some reason, you took great offence to what was a very innocuous reply from me concerning your assertion that the Iraqi health care system had " collapsed." I wasn't trying to embarrass you, nor could you have come to such a conclusion from my post. Moreover, I conceded the very point you're trying to make right now in an attempt to save face, when I stated that " ....... at various times, the health care system in Iraq has been degraded by the instability in certain regions."

Here's my response in full, I think you need a reminder of who started this nonsense:

"This was the situation Rostrup noted just after the fall of Saddam. I won't argue that in some parts of Iraq since then, at various times, the health care system in Iraq has been degraded by the instability in certain regions. I just find the generalization that the health care system has " collapsed," to be inaccurate."

I'd hate to see how stubborn you'd be if I'd actually said something that did constitute a wounding of your pride.

The words you ignore from Rostrup's statemnt are "at thi point in time".

I think it's time to admit we aren't going to charge each others minds here.

By the way, you didn't "wound my pride". You offended me with your lack of intellectual honesty.

Ian:

I don't think it's a matter of admitting " .... we aren't going to charge each others minds here."

This discussion isn't about a genuine changing of minds. You simply don't like admitting you're wrong. I don't either, but usually I'm capable of conceding the obvious, when I'm proven wrong.

You're not there yet. When you are, then you can lecture me about "intellectual honesty."

(As the comments thread this was posted to was closed before anyone replied, I'll pose these questions again - and will answer them myself if no one else wishes to. That's if Tim doesn't close this thread before I can do so!)

Three-part New Year's Quiz:

How many Iraqi civilians were violently killed during (not subsequent to) the March - April 2003 invasion of Iraq?

(a) "Significantly" more than 3,240, according to whom, based on what?

(b) "Up to" 7,350, according to whom, based on what?

(c) "About" 15,500, according to whom, based on what?

Bonus question: Which of these has the widest margin of error, and why?

Answers:
(a) The Associated Press:

"The AP count was based on records from 60 of Iraq's 124 hospitals - including almost all of the large ones - and covers the period between March 20, when the war began, and April 20, when fighting was dying down and coalition forces announced they would soon declare major combat over. AP journalists traveled to all of these hospitals, studying their logs, examining death certificates where available and interviewing officials about what they witnessed."The count is still fragmentary, and the complete toll - if it is ever tallied - is sure to be significantly higher."

(b) Iraq Body Count, whose database

"only includes individual or cumulative deaths as directly reported by the media or tallied by official bodies (for instance, by hospitals, morgues and, in a few cases so far, NGOs), and subsequently reported in the media...[and which incorporates] up to 7,350 deaths which resulted from coalition military action during the "major-combat" phase prior to May 1st 2003."

(c) "The Lancet study" - proper title: Mortality before and after the 2003 invasion of Iraq: cluster sample survey, Roberts et al.) -

- wherein, on the authors' "conservative assumptions," every recorded death represents about 3,100 others. Five violent deaths were recorded by Roberts and colleagues for the two-month period from March 1st to April 30th (see Figure 2 on page 5).

All three counts have been unequivocally described by their authors as being of civilians. I leave the "bonus question" on error margins open for others to ponder.

Mike;

"It's difficult to make the argument that the 7 men and children who died from heart attack and stroke can be attributed to invasion, when we have 3 women dying before invasion of the same cause, and only 1 after."

No it isn't.

There's considerable evidence that psychological stress can increase the incidence of sudden cardiac death. This is particularly the case very sudden and severe events, such as an earthquake or a war. You can search PubMed for the references yourself, if you're interested.

While I'm writing, does anyone out there have an explaination of the sampling strategy used? As I requested above.

Mike;

"It's difficult to make the argument that the 7 men and children who died from heart attack and stroke can be attributed to invasion, when we have 3 women dying before invasion of the same cause, and only 1 after."

No it isn't.

There's considerable evidence that psychological stress can increase the incidence of sudden cardiac death. This is particularly the case very sudden and severe events, such as an earthquake or a war. You can search PubMed for the references yourself, if you're interested.

While I'm writing, does anyone out there have an explaination of the sampling strategy used? As I requested above.

Mike;

"It's difficult to make the argument that the 7 men and children who died from heart attack and stroke can be attributed to invasion, when we have 3 women dying before invasion of the same cause, and only 1 after."

No it isn't.

There's considerable evidence that psychological stress can increase the incidence of sudden cardiac death. This is particularly the case very sudden and severe events, such as an earthquake or a war. You can search PubMed for the references yourself, if you're interested.

While I'm writing, does anyone out there have an explaination of the sampling strategy used? As I requested above.