Psyche, Science, and Society

Comments on the reception of new NEJM Iraq mortality study

January 12th, 2008

I have just posted the following comments on the new NEJM study of Iraq mortality on the Media Lens Message Board, in response to heated criticism of the new study:

 I don’t think this is fair. The NEJM study is another attempt to do something very difficult: assess the consequences of the war and occupation in a situation of extreme violence. I notice that Les Roberts was fairly positive, while raising a number of important issues. [One version of Les' thoughts can be read  [url=http://psychoanalystsopposewar.org/blog/2008/01/10/les-roberts-on-new-iraq-mortality-study/]here[/url]] I think we should follow Les’ example here.

While there are many issues with the new study, there is no fatal flaw.

I think the authors assess deaths due to violence because they believe, rightly or wrongly, that they can estimate this figure more accurately than total excess mortality. Violent deaths do not rely upon an accurate estimate of prewar mortality several years earlier, while excess mortality estimates do. This is something in which the NEJM study clearly fails. This failure does suggest, as  Les suggests, that the NEJM study is an undercount. Other problems are, as Cockburn points out, the steady rate of mortality in the NEJM study.

But Cockburn attributes nefarious motivations to the fact that Iraqi interviewers were sent to Amman for training. I will bet that this was so that they could be trained by WHO staff. Remember, at this time Les Roberts and, I believe, Gilbert Burnham, went to Amman when conducting L2 and conducted the data analyses there. The reason was the same: they felt it was too dangerous for foreigners to go into Iraq.

As for the ORB study, I was impressed when it came out. but the absence of any publication of methodological details, much less their failure to post the additional results they promised for early October cast doubt upon the study. Until they publish more, it can’t be taken as meaning much of anything, alas.

I’m afraid we’re in danger of falling into a dangerous trap of defending heartily studies whose the results we like and attacking those whose results we dislike. I teach my research students that we should subject studies confirming our prior beliefs to extra scrutiny while being careful not to search mightily for methodological flaws in those studies we don’t agree with. Otherwise, we learn nothing.

If Les welcomes this study, while examining its weaknesses, I suggest we should as well. Examining violent mortality in Iraq is extremely difficult. we may never know what the true figure is. As of summer 2006, it was most likely somewhere between 150,00 and 650,000. By now, it is probably somewhere between 250,00 and 1.2 million.  In any terms, that is truly horrifying and a humanitarian catastrophe. We should work to get that message out. To fight NEJM vs Lancet will only deflect the  message and work to the right’s advantage. Let’s not give them that advantage.

I suspect there will be responses at the Media Lens Message Board. Go there and read them.

Entry Filed under: Iraq, Mortality, Public Health, Research Methods, Violence