Left I on the News

The World Health Organization has now published an article in the prestigious New England Journal of Medicine new study on deaths in Iraq, and the study has gotten the official stamp of approval from the Washington Post. Here's their lead:

A new survey estimates that 151,000 Iraqis died of violence in the three years following the U.S.-led invasion of the country. Roughly nine of 10 of those deaths were a consequence of U.S. military operations, insurgent attacks and sectarian warfare.

The survey, conducted by the Iraqi government and the World Health Organization, also found a 60 percent increase in nonviolent deaths -- everything from childhood infections to kidney failure -- during the period.

I haven't had time to do a complete analysis, but right off the bat you can see some problems. First, note that both the news articles and the report itself quantify only the 151,000 deaths from "violence." This does make it clear how inaccurate the methodology of Iraq Body Count (its much lower numbers relying on published media reports, and counting "civilians" only) is, but it's very misleading with respect to the earlier Johns Hopkins studies. Why? Because the Johns Hopkins study was a study of "excess deaths," not deaths by violence only. Unfortunately, although the WHO study says they "found a 60 percent increase in nonviolent deaths," they don't quantify how many that amounts to, so comparing its 151,000 total to the Johns Hopkins study is difficult indeed.

Also, note the March June 2006 end date. Nothing wrong with that, the study had to end sometime, but there's no attempt, either in the press or in the study itself, to project from that number or even to mention it. 21 18 months have passed since March June 2006, and a lot more Iraqis have died (indeed, official numbers have said that last year saw an even higher Iraqi death toll than the ones that preceded it). Clearly, then, the "151,000" number, which is now going to assume the role of gospel, is starting out incorrect, even if it was correct as of March June 2006.

But the biggest problem is the non-violent death problem. There seems to be an idea that only violent deaths "count," as if people dying from poor public health conditions, poor nutrition, or poor health care are somehow less dead, or as if the increase in their numbers is any less attributable to the invasion. Consider Table 3 from the study. For all ages, subtracting out the violent component gives 3.07 deaths per 1000 person-years from disease and other non-violent causes before the invasion, and 4.92 after, a 60% increase. Violent deaths went from 0.1 to 1.09. I'm no statistician, but even though that's a 10-fold increase, the absolute increase of 1.85 deaths per 1000 person-years from non-violent causes would seem to be 70% larger. So if there were 151,000 additional violent deaths by March 2006, my crude calculation (which I am very willing to have corrected by a real statistician) gives 256,000 deaths from non-violent causes, for a total of 407,000 Iraqis dead as a result of the invasion by March June 2006. That's three years of data, which means it's 11,300/month. Add an another 21 18 months and that's another 238,000 203,400 people, for a grand total of 645,000 610,000, more than four times higher than the number you're now going to be hearing bandied about in the corporate media.

One hell of a lot of people. Or, to be blunt about it, former people. They're dead now.

P.S.: Shall we start a poll as to when the first time a reporter will ask George Bush about these numbers? I'll place my bet on "never."

Update: First misinterpretation: Democracy Now! this morning reports that "a new study shows that 151,000 Iraqis have died." No, the new study shows that 151,000 Iraqis have been killed by violent means through June 2006.

Second misinterpretation, from The New York Times: "W.H.O. Says Iraq Civilian Death Toll Higher Than Cited." No, the WHO study has nothing to do with "civilians," it has to do with "Iraqis." The only appearance of the word "civilians" in the NEJM paper is in conjunction with mentions of Iraq Body Count, which does count only "civilians." The Times further confuses deaths with violent deaths, reporting that "The World Health Organization said its study...indicated with a 95 percent degree of statistical certainty that between 104,000 and 223,000 civilians had died." It did no such thing. It indicated that many Iraqis had died from violence. Quite a different thing.

Small correction: Just caught myself in one math error. Somehow I thought that the study ended in March, 2006; it actually ended in June, 2006. That means I added in three too many more months of deaths. I should have added in 18 months, for a total of 610,000 deaths, not 645,000.

Another update: I just checked the Democracy Now website, and it's worse than I wrote above. Like the NY Times, they report (inaccurately on three counts!): "A new study of the civilian death toll since the U.S. invasion of Iraq has put the number of Iraqi deaths at 151,000." Not "civilians," not the "death toll," and not "since the invasion". This from the most progressive news organization with reasonably mass distribution in the entire country.