An Interview with EPIC Advisor Richard Garfield
In early November 2004, Michelle Suwannukul talked with Richard
Garfield of Columbia University, one of the lead authors of a recent
Iraq mortality study published in leading public health journal, The
Lancet. Dr. Garfield is also part of EPIC's board of advisors and a
member of our speakers bureau.
Compared to the previous wars and in particular the Gulf War in 1991, how does the most recent Iraq war related deaths compare?
The most recent Iraq War was a much larger war involving much more
territory and fire power [but] caused a more limited number of deaths.
What we have seen since is a much different story. US occupations from
WWII forward [we have seen] virtually no casualties once the occupation
began both amongst civilians and amongst military. This started to
change particularly with Afghanistan and Iraq [where it is really]
different from our pattern of historic occupation over the last sixty
years. Most dramatically for me, this is the first time the US has
[been involved in a war where] there have been more casualties among
our troops in a period after major conflict ended than during the
period of conflict.
The perception is that there weren’t that many soldiers that died in
the Gulf War of 1991. But because nowadays [wars] can be very high-tech
and very rapid, it is illuminating to look at the rate of mortality of
10,000 troops per year and make it a yearly rate. What you can see
is that we did indeed have a rapid reduction in mortality through
the twentieth century up through Viet Nam and then the trends started
to change. In the period of major conflict the rate of mortality was
three times higher last year than in the war in 1991. More important
because it is a much longer period of time in terms of mortality among
Iraqi security forces.
When speaking of war related deaths, what specifically are you
referring to? And what are the reasons behind choosing one group’s
deaths over another as an indicator?
I’m focusing, so far, on deaths among troops because that’s where we
know something. We have excellent information on deaths of our troops.
When you leave that arena and go to the status of life of the civilian
population which is the overwhelming majority of people [physically]
involved in the geographically important area, that’s where the
information starts to break down very rapidly. Our best single source
of count data on mortality of Iraqis comes from the morgue in Baghdad.
And here you see that deaths recorded in the Baghdad morgue were,
for a long period, around 200 per month. Recording fell apart during
the war and then you see right after that a rapid rise in the rate of
mortality. This is the experience among those who are recording mortal
events in Baghdad. Most of these excess deaths are due to violence,
most of them are gunshot wounds.
Considering that many of these deaths are recorded from the
morgue, how do you account for deaths where the bodies never get to the
morgue? For example, air strikes don’t really have reporters on the
ground to record the mortal events, how do you account for these?
We also have pretty good data on mass bombing casualties where you
see that even in the first months after the war there were none.
It picked up in the summer and has resulted in an increasing number
of deaths among civilians [month by month]. These still are ‘tips of
the iceberg’ of our ability to have a sense of what is going on in
terms of mortality rates amongst civilians [and we combine this] week
by week with the data we have on deaths among coalition soldiers.
[The first week after major combat ended] This was the period when
we accumulated more deaths among soldiers than had occurred during the
period of major combat and deaths among civilians was 5 times higher.
This is the data we get from press reports, from Iraqi Body Count
project which combines information from 38 fights where the press
report a specific incident, how many people, where it occurred, and
anything that can be confirmed. That is what we’ve had until now, it
is [was] by far our best indicator for mortality amongst
civilians. We see that the pattern of mortality is fairly similar to
those among troops, rising week by week but always at least about 5
times higher. This is among our better ‘tip of the iceberg’ indicators
because we only have these types of indicators (those accounts when the
press was present and actually reported in an international source). We
are lacking tremendously in information on the experience of Iraqis
alone. Until now, Washington and 10 Downing St. (Prime Minister’s
office) in London have ignored the Iraqi Body Count project as a kind
of strange, unreliable, and ‘who knows what they’re doing’ kind of
Isn’t US targeting good enough that most violent deaths will be of insurgents rather than civilians?
It may be true, when targeting is done well, that most Iraqis killed
will be insurgents. Close to half of all violent deaths found in the
study occurred among adult males. Since adult males are only about a
quarter of the population, their chances of dying were lower than
women or children. However, the remaining half dying violent deaths
were not adult males, and therefore less likely to be insurgents. The
people reported to have died had names, ages, and family relations. Our
veracity check of comparing the data to death certificates were
consistent with the verbal information gathered.
Could you talk about the methodology used in the Lancet study?
We gathered with a group of six Iraqis in the month of September
, we carried out a mortality survey with a multi-stage cluster
sample survey which covered the entire country of Iraq. During the
period of September, we reached all 33 clusters that were chosen
randomly to be in the sample which collected information on 988
households, 30 households per cluster and identified who had been
living there at any time for more than a two month period over the last
two years starting the year before the war last year.
What were the findings and what was your impression of what you found?
We knew that the Iraqi Body Count project data, which shows the
mortality rate being 5 times higher than those of the troops, we didn’t
know how much was below that. We knew that these points were important
and we knew that we really didn’t know that much. Nonetheless, we were
really surprised at what we found.
First of all, very few people refused or were unable to take part in
the sample, to our surprise most people had death certificates and we
were able to confirm most of the deaths we investigated. Here is the
One of the first clusters we had was in the town of Fallujah. [The
deaths] were so much higher than in other parts of the country so we
left it out of our analysis. But if you take just deaths that occurred,
excluding areas characteristic of high conflict, we saw an excess
mortality rate of about 50%, which is to say that all mortality rose
50% in the year after the war compared to the year before it.
In areas of Iraq, with the exception of the North, all had a rise in
the mortality rate and most were due to violence. Real change was in
deaths due to violence.[The majority of the 57,600 deaths due to
violence was attributed to air assaults.]
Basically, the Iraqi Body Count project had accumulated data by
counting death events for about ¼ of all violent deaths. ¾ went below
the radar and hasn’t been counted and a little more than ½ of all
violent deaths occurred during aerial attacks which explains why there
hasn’t been any data because if there weren’t ground troops involved
then there wasn’t a reporter there anticipating and recording data.
We identified an excess of as many as 100,000 deaths [a range
was given of 8,000-194,000 so the 'true' value could be below
or above the mean of 98,000] in the period since the war. There
are some changes relating to conditions of life, infrastructure, and
access to medical care but the overwhelming change in mortality in the
year 2004 has to do with violence.
This was published in the Lancet, this was kind of remarkable
research done in September and we were writing the report as we were
traveling back almost 3 weeks after the field work was done.
Read the Lancet Report>>>