When a controversial study was released in The Lancet last October, suggesting that by that time some 655,000 people had died (over and above normal mortality rates) in Iraq since the war had started, Republican pundits went crazy. That was more than 10 times any official estimate. Even George Bush (who certainly studied statistical epidemiology at Yale) was quick to criticize the “poor methodology” of the Johns Hopkins team.
Now, in response to the real criticism of university professionals, the researchers are responding by releasing raw data to some–but not all–of their peers. The fire storm continues.
The authors of the study (lead by Gilbert Burnham of Johns Hopkins) got their results by extrapolating from the results of door-to-door surveys, and validating some of those results by cross checking with death certificates. They employed local resident and trained them to take the surveys.
Critics like Michael Spagat, from Royal Holloway, University of London, suggests there is a bias in their methods, since the quick survey results may have relied on main street homes–more susceptible to IEDs and car bombs than back street or more rural settings. There is also the issue of some raw data being destroyed in the chaos of Baghdad. These factors have led some blogs on the right to join the chorus accusing the researchers of ginning up the numbers for political reasons:
Now lead researcher Gilbert Burnham’s team is releasing some of the raw data stripped of information that might reveal identities–but only to groups that they consider “qualified.” They define that as those “with expertise in biostatistics and epidemiology…and ‘without publically stated views that would cause doubt about their objectivity in analyzing the data.'” [Quoted from Science, 316:20 April, 2007]
Spagat was not allowed access to the data, because the research team suggested he “would not meet the criteria by multiple measures.” The feud was featured prominently in this week’s Science.
Many people believe that the halls of scientific research are normally quite quiet and staid; not so with this epidemiological spat. Accusations have continued to go back and forth, fired by the vehemence of political fervor. Without access to the original documents, or the ability to interview the brave employees who did the street surveys (for their own protection) the issue may never be truly resolved by scientific standards.
In the midst of the war of words, the essential message of the study seems to have been lost–the true death toll in Iraq is many times what the Bush administration is willing to admit. Neither have they discussed the growing number of refugees in Jordan, Syria and other nations where they have been admitted.
The situation parallels the political games being played with death rates and injury rates of American service persons in Iraq, where the distinction between a death that is “combat related” and one that is not seems to be which seat you are sitting in then an IED hits. (The backlog of over 600,000 potential combat-related injuries in the VA is another sign of ongoing effects to deceive.)
Does it really matter whether the number is 655,000 or 555,000? Any number that is an order of magnitude greater than (statistician) Bush will admit is a human tragedy beyond imagination.
