Study Revises Iraq's Death Toll

By John Bohannon
ScienceNOW Daily News
9 January 2008

Between 104,000 and 223,000 Iraqis died violent deaths between the start of the U.S.-led invasion of Iraq in March 2003 and June 2006. That's the conclusion of a study by the Iraqi government and the World Health Organization (WHO) published online today in The New England Journal of Medicine. The new estimate throws fresh doubt on a controversial study led by researchers at Johns Hopkins University in Baltimore, Maryland, published in The Lancet in 2006, that estimated a death toll four times higher (Science, 20 October 2006, p. 396).

Between 2006 and 2007, Iraqi surveyors led by WHO epidemiologist Mohammed Ali knocked on the doors of 9345 Iraqi homes. The addresses were chosen randomly within about 1000 neighborhoods and villages across the country. Besides basic information about the household, interviewers asked about illnesses and death among occupants since the beginning of 2002.

This survey, like the earlier Lancet study, found that violence became a leading cause of death in Iraq after the March 2003 invasion. The WHO team used the same method of extrapolation to calculate total death rate as the Lancet group did, but their average estimate of violent deaths is 151,000--far lower than the Lancet's 601,000. When the WHO surveyors were assigned a cluster of households in a neighborhood deemed too dangerous to visit, they used data from Iraq Body Count, a London-based database of media-reported Iraqi casualties whose numbers are considered a floor. "Some homes could not be visited because of high levels of insecurity," says Iraqi health minister Salih Mahdi Motlab Al-Hasanawi, who was not an author. "Nonetheless, the survey results indicate a massive death toll since the beginning of the conflict."

Researchers say any attempts to calculate death tolls during a war will have limitations, but they are impressed by the new study. "There are inherent biases in a household survey like this one," says John Brownstein, an epidemiologist at Harvard Medical School in Boston, particularly when the total population size is not well-known. "But their approach was rigorous, and they accounted for all reasonable criticisms."

Brownstein declined to comment on the Lancet study, noting only that the WHO study's sample size--five times larger than that of the Lancet study--makes it "more robust, all things being equal." But some researchers are not as forgiving. "This study is a slaughter of the Lancet study," says Michael Spagat, an economist at Royal Holloway, University of London, in the U.K. "Serious problems with that study have been evident for a long time, but this confirms it."