In an update of a two-year-old survey that sparked wide disagreement, Johns Hopkins researchers now estimate that more than a half-million Iraqis have died as a result of the U.S.-led invasion and its bloody aftermath.
Reporting this week in the online edition of The Lancet, a leading British medical journal, the researchers estimated that 654,000 more Iraqis died of various causes after the invasion than would have died in a comparable period before.
The scientists attributed 600,000 of those deaths to acts of violence.
Gunshots emerged as the leading cause of death, accounting for 56 percent of the total. Airstrikes, car bombs and other explosions each accounted for 13 percent to 14 percent. Almost 60 percent of the deaths were among males 15 to 44.
"In this conflict, like all other recent conflicts, it's the population that bears the consequences," said Dr. Gilbert Burnham, lead author and co-director of the Center for Refugee and Disaster Response at the Johns Hopkins Bloomberg School of Public Health.
"To put these numbers in context, deaths are occurring in Iraq now at a rate more than three times that from before the invasion of March 2003," he said.
With deaths increasing at that pace, Burnham said, the crisis in Iraq qualifies as a "humanitarian emergency," a public health term applied to situations in which populations die at twice or more the usual rate.
The Pentagon declined to comment on the study, saying that officials had not yet read it. Army Lt. Col. Christopher Garver, a spokesman for the Multi-National Force in Iraq, said the force was "interested" in the study but doesn't normally comment on Iraqi deaths.
"We're obviously trying to reduce the amount of violence of all kinds in Iraq, but it's really a government of Iraq issue," he said.
In August, the Defense Department reported to Congress that the number of civilian casualties had increased sharply, without estimating total deaths.
In 2005, however, the Bush administration estimated that 30,000 civilians had died. A report by the Los Angeles Times in June, based on statistics from the Iraqi Health Ministry and the Baghdad morgue, put civilian deaths at 50,000.
To estimate deaths for this week's report, the Hopkins group recruited Iraqi doctors to conduct household surveys in 47 neighborhood clusters across Iraq that contained 1,849 households and 12,801 people. The doctors asked family members to report births, deaths and the movement of people into and out of their households.
When people reported deaths, researchers asked them about the cause and obtained death certificates in 92 percent of cases. The data were then projected onto the entire nation, about 26 million people.
The Hopkins estimate is many times higher than any other group's, including Iraq Body Count, a Web-based organization that put the death count at 48,693 yesterday. Members of that group, which criticized Hopkins' earlier estimate as wildly inflated, were unavailable for comment.
Unlike the Hopkins study, Iraq Body Count has based its estimates on reports from morgues and the news media.
Michael O'Hanlon, a Brookings Institution scholar who criticized the last Hopkins study, said the method used by the Hopkins researchers this time was seriously flawed.
"The numbers are preposterously high," he said. "Their numbers are out of whack with every other estimate."
Burnham, however, said that no one else has done a population-based survey of deaths in Iraq. "There are people who have taken numbers reported from various facilities and so forth, but statistically you can't [extrapolate] from various morgue and newspaper reports to a national figure," he said.
He said the group employed standard epidemiological methods used to estimate deaths from calamities ranging from natural disasters such as Hurricanes Andrew and Katrina to the bloody war in the Congo.
In its earlier study, published in October 2004, the Hopkins group estimated that close to 100,000 people had died in the 17 months after the U.S.-led invasion in March 2003. Critics attacked that study, in part, because of a wide "confidence interval," a range within which the true number exists.
In that case, the interval was 8,000 to 194,000 deaths, though researchers said 100,000 was most likely the actual number.
To avoid that problem this time, researchers increased the number of neighborhood clusters from 33 to 47. Additionally, Burnham said, by the time the second survey was conducted, hostilities had spread more widely through the country, reducing the possibility that the overall projection was skewed by pockets of extreme violence.
The researchers now say that Iraqi deaths totaled between 392,000 to 942,000. Their best estimate is 654,000.
In the earlier study, Iraqis told interviewers that about a third of all deaths were caused by coalition forces. Since then, the proportion has dropped to about a quarter - although the absolute number of deaths has risen in each year of the conflict.
From June 2005 to June 2006, residents attributed about 30 percent of deaths to "other" causes, including sectarian or criminal activity, while 44 percent were reported as "unknown."
The researchers reasserted their earlier call for an independent body to assess deaths and monitor compliance with the Geneva Conventions.
In a commentary, Lancet editor Richard Horton denounced the coalition strategy but said that "absolute despair would be the wrong response."
"Instead, the disaster that is the West's current strategy in Iraq must be used as a constructive call to the international community to reconfigure its foreign policy around human security rather than national security, around health and well being in addition to the protection of territorial boundaries and economic stability," he said.
Sun reporter Jonathan Rockoff and the Los Angeles Times contributed to this article.