A new study by public health researchers estimates that up to 600,000 Iraqi people — nearly 1 in 40 — have died violently since the American-led invasion of the country in March 2003.

The estimate, which far exceeds figures compiled by the United Nations and the Iraqi Government, is the second made by a group of American and Italian researchers and used a sampling of nearly 2,000 households across Iraq to extrapolate a total number of violent deaths, be they caused by crime, the US-led coalition or sectarian strife.

The first report, issued in October 2004 by a team led by Les Roberts of the Johns Hopkins Bloomberg School of Public Health, estimated that 100,000 people had been killed in the first year of the war. The study was criticised for its narrow sample and wide margin of error.

The new study, published in the online edition of The Lancet, the British medical journal, also accepts a broad range of error, with its lead author, Gilbert Burnham, also of Johns Hopkins, saying the true figure could lie anywhere between 426,369 to 793,663.

It estimated that a total of 654,965 more Iraqis had died as a consequence of the war than "would have been expected in a non-conflict situation". Of those, 601,000 it was said had died directly of violent causes, including gunfire, car bombs, air strikes and other explosions. The rest had suffered from a general decline in healthcare and sanitary standards due to failing water supplies, sewerage and electricity supply.

The survey drew swift and senior rebuttals. When asked about the study at a White House news conference, President George Bush said: "The methodology is pretty well discredited."

He added: "I do know that a lot of innocent people have died and that troubles me, and it grieves me. And I applaud the Iraqis for their courage in the face of violence."

The researchers defended their methods, which replicate those used to estimate the death toll in humanitarian emergencies such as Darfur, claiming that studying the mortality rate of a sample of families across Iraq is at least as accurate as relying on casualty figures issued by morgues, hospitals and the Iraqi Government.

According to a report in today's New York Times, the researchers maintain that their study reflects the larger breakdown of order across Iraq and reflects the turbulence outside Baghdad, which dominates press and official reports about the progress of the war.

"We found deaths all over the country," Dr Burnham told the newspaper, adding that Baghdad was an area of medium violence compared to the provinces of Diyala and Salahuddin, north of the capital, and Anbar to the west, which all had higher death rates.

The study found that up to 15,000 people are dying violently every month in Iraq, a level that surpasses by far the most recent UN estimates.

Last month the UN Assistance Mission in Iraq said that 3,009 civilians had died violently in August, down from 3,590 in July, two of the worst months of the war so far. More than 5,000 of the deaths were reported in Baghdad.

The US military does not keep an official count of the civilian casualties in Iraq, but according to its latest report to Congress, around 120 Iraqis, including police officers and soldiers, died every day in August, a total of 3,600, up from 26 a day, or 800 per month, in 2004. The John Hopkins figures also tower over the running totals maintained by the Iraq Body Count, an independent group that monitors media reports to estimate the numbers of Iraqi dead. The group's current total stands at 48,000.

The Iraqi Government, meanwhile, has sought to take control of the compilation of mortality statistics. Baghdad's central morgue, until now the main source of information for violence in the capital, was prohibited from issuing its own information last month.

Today, Adel Mohsin, the Iraqi Deputy Health Minister, cast doubt on the estimate, saying: “I think it’s a bit exaggerated... I’d say we are now averaging about 2,000 to 3,000 maximum a month killed, which would be 36,000 a year.”

Mr Mohsin did say that 50,000 Iraqis could have lost their lives because of the parlous state of the country's hospitals and infrastructure: “Obviously if we compare the standard of treatment to countries like the UK, we have lost a lot of people," he said.

The authors of the Johns Hopkins study chose 1,849 families from 47 districts across Iraq — chosen for their geographical location and population size, rather than level of violence — and found that the death rate among the 12,801 people they studied was 13.3 deaths per 1,000 people since the war began. That rate compared to an estimate of 5.5 per 1,000 under Saddam.

The researchers found that gunfire caused 56 per cent of the deaths directly attributable to violence, with air strikes and car bombs accounting for a further 13 to 14 per cent.

Commenting on the results, Dr Richard Horton, editor of The Lancet, said: "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."

Times Recommends

• Tehran cannot live up to its rhetoric
• The new unicycle that will ‘change the world’
• We cannot wish away Iran’s bomb