Important note: Images and videos posted on this
website are very graphic. Viewers discretion is strongly advised!
The Lancet: Report
Mortality Before And After The 2003 Invasion
Cluster Sample Survey
Les Roberts, Riyadh Lafta, Richard Garfield, Jamal
Khudhairi, Gilbert Burnham
In March, 2003, military forces, mainly from the USA and the UK,
invaded Iraq. We did a survey to compare mortality during the period of 14·6 months before the invasion with the 17·8 months
A cluster sample survey was undertaken throughout Iraq during September,
2004. 33 clusters of 30 households each were interviewed about household composition, births, and deaths since January, 2002.
In those households reporting deaths, the date, cause, and circumstances of violent deaths were recorded. We assessed the
relative risk of death associated with the 2003 invasion and occupation by comparing mortality in the 17·8 months after the
invasion with the 14·6-month period preceding it.
The risk of death was estimated to be 2·5-fold (95% CI 1·6–4·2)
higher after the invasion when compared with the preinvasion period. Two-thirds of all violent deaths were reported in one
cluster in the city of Falluja. If we exclude the Falluja data, the risk of death is 1·5-fold (1·1–2·3) higher after
the invasion. We estimate that 98 000 more deaths than expected (8000–194 000) happened after the invasion outside of
Falluja and far more if the outlier Falluja cluster is included. The major causes of death before the invasion were myocardial
infarction, cerebrovascular accidents, and other chronic disorders whereas after the invasion violence was the primary cause
of death. Violent deaths were widespread, reported in 15 of 33 clusters, and were mainly attributed to coalition forces.
Most individuals reportedly killed by coalition forces were women and children. The risk of death from violence in the
period after the invasion was 58 times higher (95% CI 8·1–419) than in the period before the war.
Making conservative assumptions, we think that about 100 000
excess deaths, or more have happened since the 2003 invasion of Iraq. Violence accounted for most of the excess deaths and
air strikes from coalition forces accounted for most violent deaths. We have shown that collection of public-health information
is possible even during periods of extreme violence. Our results need further verification and should lead to changes to reduce
noncombatant deaths from air strikes.
The number of Iraqis dying because of conflict or sanctions since the 1991 Gulf war is uncertain.[1,2
Claims ranging from a denial of increased mortality[3–7
] to millions of excess deaths[8
have been made. The Coalition Provisional Authority and the Iraqi Ministry of Health have identified the halving of infant
mortality as a major objective.[9
] In the absence of any surveys, however, they have relied on Ministry of
Health records. These data have indicated a decline in young child mortality since February, 2001, but because only a third
of all deaths happen in hospitals, these data might not accurately represent trends.[10
] No surveys or censusbased
estimates of crude mortality have been undertaken in Iraq in more than a decade, and the last estimate of under-five mortality
was from a UNICEF-sponsored demographic survey from 1999.[11,12
] Morgue-based surveillance data indicate
the postinvasion homicide rate is many times higher than the preinvasion rate. In Baghdad, a city of 5 million people, 3000
gunshot-related deaths happened in the first 8 months of 2004.[13
] One project has kept a running estimate
of press accounts of the number of Iraqi citizens killed by coalition forces: at present, the estimated range is 13 000–15
). Aside from the likelihood that press accounts are incomplete, this source does not record deaths that are the indirect
result of the armed conflict. Other sources place the death toll much higher.[14
] In a recent BBC article
decrying the lack of a reliable civilian death count from the war in Iraq, Ken Roth of Human Rights Watch purports that it
will not be possible “to come up with anything better than a good guess at the final civilian cost”.[14
In the present setting of insecurity and limited availability of health information, we undertook a nationwide survey to estimate
mortality during the 14·6 months before the invasion (Jan 1, 2002, to March 18, 2003) and to compare it with the period from
March 19, 2003, to the date of the interview, between Sept 8 and 20, 2004.
We designed the cross-sectional survey as a cohort study, with every cluster of households essentially matched to itself
before and after the invasion of March, 2003.
Assuming a crude mortality rate of 10 per 1000 people per year, 95% confidence, and 80% power to detect a 65% increase
in mortality, we derived a target sample size of 4300 individuals. We assumed that every household had seven individuals,
and a sample of 30 clusters of 30 households each (n=6300) was chosen to provide a safety margin. We selected 33 clusters
in anticipation that 10% of selected clusters would be too insecure to visit.
We obtained January, 2003, population estimates for each of Iraq’s 18 Governorates from the Ministry of Health.
No attempt was made to adjust these numbers for recent displacement or immigration. We assigned 33 clusters to Governorates
via systematic equal-step sampling from a randomly selected start. By this design, every cluster represents about 1/33 of
the country, or 739 000 people, and is exchangeable with the others for analysis. Most communities visited consisted of fewer
than 739 000 people. Thus, when referring to a specific cluster by name, this group of 30 households is representing 1/33
or 3% of the country, which may extend beyond the confines of that village or city.
During September, 2004, many roads were not under the control of the Government of Iraq or coalition forces.
Local police checkpoints were perceived by team members as target identification screens for rebel groups. To lessen
risks to investigators, we sought to minimise travel distances and the number of Governorates to visit, while still sampling
from all regions of the country. We did this by clumping pairs of Governorates. Pairs were adjacent Governorates that the
Iraqi study team members believed to have had similar levels of violence and economic status during the preceding 3 years.
The paired Governorates were:
Basrah and Missan, Dhi Qar and Qadisiyah, Najaf and Karbala, Salah ad Din and Tamin, Arbil
and Sulaymaniya, and Dehuk and Ninawa.
All clusters were assigned to Governorates without regard to any security considerations. Then, for the six sets of paired
Governorates, a second phase of cluster assignment took place. The populations of the two Governorates were added together,
and a random number between 0 and the combined population was drawn. If the number chosen was between 0 and the population
of the first Governorate, all clusters previously assigned to both clusters went to the first. Likewise, if the random number
was higher than the first Governorate population estimate, the clusters for both were assigned to the second. Because the
probability that clusters would be assigned to any given Governorate was proportional to the population size in both phases
of the assignment, the sample remained a random national sample. This clumping of clusters was likely to increase the sum
of the variance between mortality estimates of clusters and thus reduce the precision of the national mortality estimate.
We deemed this acceptable since it reduced travel by a third. Table 1 presents cluster groupings and figure 1 shows the location
Estimated populations (millions)
Clusters initially assigned at random
Clusters visited after grouping process
(Numbers in parentheses denote pairings of Governorates.)
Table 1: Estimated populations of Governorates (January, 2003) and assignment of clusters
We assigned clusters to individual communities within the Governorates by creating cumulative population
lists for the Governorate and picking a random number between one and the Governorate population. Once a town, village, or
urban neighbourhood was selected, the team drove to the edges of the area and stored the site coordinates in a global positioning
system (GPS) unit. We assumed the population was living within a rectangle, with the dimensions corresponding to the distances
spanned between the site coordinates stored in the GPS unit. The area was drawn as a map subdivided by increments of 100 m.
A pair of random numbers was selected between zero and the number of 100 m increments on each axis, corresponding to some
point in the village. The GPS unit was used to guide interviewers to the selected point. Once at that point, the nearest 30
households were visited.
The study teams included at least a team leader and one male and one female interviewer. Five of
the six Iraqi interviewers were medical doctors. All six were fluent in English and Arabic. All interviewers participated
in the revisions and two rounds of fieldtesting of the questionnaire. Data were recorded in English.
Households were informed about the purpose of the survey, were assured that their name would not
be recorded, and told that there would be no benefits or penalties for refusing or agreeing to participate. We defined households
as a group of people living together and sleeping under the same roof(s). If multiple families were living in the same building,
they were regarded as one household unless they had separate entrances onto the street. If the household agreed to be interviewed,
the interviewees were asked for the age and sex of every current household member. Respondents were also asked to describe
the composition of their household on Jan 1, 2002, and asked about any births, deaths, or visitors who stayed in the household
for more than 2 months. Periods of visitation, and individual periods of residence since a birth or before a death, were recorded
to the nearest month. Interviewers asked about any discrepancies between the 2002 and 2004 household compositions not accounted
for by reported births and deaths. When deaths occurred, the date, cause, and circumstances of violent deaths were recorded.
When violent deaths were attributed to a faction in the conflict or to criminal forces, no further investigation into the
death was made to respect the privacy of the family and for the safety of the interviewers. The deceased had to be living
in the household at the time of death and for more than 2 months before to be considered a household death.
Within clusters, an attempt was made to confirm at least two reported non-infant deaths by asking
to see the death certificate. Interviewers were initially reluctant to ask to see death certificates because this might have
implied they did not believe the respondents, perhaps triggering violence. Thus, a compromise was reached for which interviewers
would attempt to confirm at least two deaths per cluster. Confirmation was sought to ensure that a large fraction of the reported
deaths were not fabrications. Death certificates usually did not exist for infant deaths and asking for such certificates
would probably inflate the fraction of respondents who could not confirm reported deaths. The death certificates were requested
at the end of the interview so that respondents did not know that confirmation would be sought as they reported deaths. We
defined infant deaths as deaths happening in the first 365 days after birth. Violent deaths were defined as those brought
about by the intentional acts of others.
For most clusters, the latitude and longitude was recorded. At the end of interviewing every 30 household
cluster, one or two households were asked if in the area of the cluster there were any entire families that had died or most
of a family had died and survivors were now living elsewhere. We did this to explore the likelihood that families with many
deaths were now unlikely to be found and interviewed, creating a survivor bias among those interviewed. Houses with no one
home were skipped and not revisited, with the interviewers continuing in every cluster until they had interviewed 30 households.
Survey team leaders were asked to record the number of households that were not home at the time of the visit to every cluster.
We tabulated data and calculated the number of births, deaths, and person-months associated with
every cluster. For every period of analysis, crude mortality, expressed as deaths per 1000 people per year, was defined as:
(number of deaths recorded/number of person-months lived in the interviewed households) X12X1000. We estimated the infant
mortality rate as the ratio of infant deaths to livebirths in each study period and presented this rate as deaths per 1000
Mortality rates from survey data were analysed by software designed for Save the Children by Mark
Myatt (Institute of Ophthalmology, UCL, London, UK), which takes into account the design effect associated with cluster surveys,
and reconfirmed with EpiInfo 6.0. We estimated relative and attributable rates with generalised linear models in STATA (release
8.0). To estimate the relative risk, we assumed a log-linear regression in which every cluster was allowed to have a separate
baseline rate of mortality that was increased by a clusterspecific relative risk after the war. We estimated the average
relative rate with a conditional maximum likelihood method that conditions on the total number of events over the pre-war
and post-war periods, the sufficient statistic for the baseline rate. We accounted for the variation in relative rates
by allowing for overdispersion in the regression. As a check, we also used bootstrapping to obtain a non-parametric confidence
interval under the assumption that the clusters were exchangeable. The confidence intervals reported are those obtained
by bootstrapping. The numbers of excess deaths (attributable rates) were estimated by the same method, using linear rather
than log-linear regression. Because the numbers of deaths from the specific causes of death were generally very small, EpiInfo
(version 3.2.2, April 14, 2004) was used to estimate the increased risk of cause-specific mortality without regard to the
design effect associated with the cluster data.
We estimated the death toll associated with the conflict by subtracting preinvasion mortality from
post-invasion mortality, and multiplying that rate by the estimated population of Iraq (assumed 24·4 million at the onset
of the conflict) and by 17·8 months, the average period between the invasion and the survey.
This study was approved by the Johns Hopkins
Bloomberg School of Public Health Committee on Human
Role of the funding source
The sponsors had no role in the
design of the study beyond requiring that the crude mortality be measured and that the portion attributable to violence be
documented, and they had no role in data collection, data analysis, data interpretation, or writing of the report. The corresponding
author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Read the Report in full
L Roberts was the lead investigator
in the field and was principally responsible for the data analysis, interpretation, and preparation of this report. R Lafta
was involved in study design, hired, trained, and oversaw the interview staff, led one of the two study teams, coordinated
all logistical aspects of the study, and had a central role in data interpretation and preparation of this report. R Garfield
issues of study design, study execution, participated in the analysis and interpretation of data and preparation
of this report, and initially organised the study team. J Khudhairi was involved in the study design, interviewer training,
and oversaw one of the two survey teams in the field. G Burnham advised on issues of study design, study execution, participated
in the analysis and interpretation of data and preparation of this report, and organised and facilitated the ethics review
process at Johns Hopkins University.
Conflict of interest statement
We declare that
we have no conflict of interest.
This survey was funded by the
Center for International Emergency Disaster and Refugee Studies, Johns Hopkins Bloomberg School of Public Health and the Small
Arms Survey in Geneva Switzerland, whose support is greatly appreciated. Special thanks to Walt Jones for swiftly facilitating
this project. Reference support was provided by the Sidney Memorial Library in Sidney, NY, USA and assistance with figure
1 was provided by Marite Jones. This work could not have been completed without a host of brave Iraqis who endured danger,
police interrogations, and the risk of being associated with foreign investigators. Many thanks to Elizabeth Johnson and Scott
Zeger of the
Johns Hopkins Bloomberg School of Public Health, Department of Biostatistics, for assistance with data analysis.
Finally, thanks to Helen Wolfson for data cleaning and tabulation and Mary Grace Flaherty for editing this manuscript.
1 Global Policy Forum, Save
the Children UK. Iraq sanctions:
humanitarian implications and options for the future. http://www.globalpolicy.org/security/sanction/
iraq1/2002/paper.htm (accessed Feb 7, 2003).
2 Campaign against sanctions on Iraq. http://www.cam.ac.uk/societies/casi/index.html (accessed Feb 7, 2003).
3 Clawson PA. Oil for food or the end of sanctions. Policywatch 1998;
4 Cordesman AH. Sanctions and the Iraqi people: the WHO report and
conflicting views. Washington, DC: Center for Strategic and International Studies, 1997.
5 Baram A. The effect of Iraqi sanctions: statistical pitfalls and
responsibility. Middle East J 2000; 54: 197–223.
6 Rubin M. Food fight. The New Republic; 2001 (June 18): 18.
7 Welch M. The politics of dead children: have sanctions against
Iraq murdered millions? Reason Magazine 2002: 52–59. http://www.reason.com/0203/fe.mw.the.shtml (accessed Oct 26, 2004).
8 Reuters News Service. Iraq says sanctions have killed 1·4 million.
(accessed Aug 11, 1998).
9 CPA. An historic review of CPA accomplishments.
http://www.cpa-iraq.org (accessed Sept 26, 2004).
10 Garfield R. Morbidity and mortality among Iraqi children from
1990–1998: assessing the impact of economic sanctions.
http://www.fourthfreedom.org/Applications/cms.php?page_id=7 (accessed Oct 26, 2004).
11 Ali MM, Shah IH. Sanctions and childhood mortality in Iraq.
2000; 355: 1851–57.
12 Ali M, Blacker J, Jones G. Annual mortality rates and excess deaths
of children under five in Iraq, 1991-98. Popul Stud 2003, 57: 217–26.
13 Berenson A. Killings surge in Iraq, and doctors see a procession
misery. New York Times, Sept 26, 2004.
14 Davis M. Counting the civilian cost in Iraq. http://news.bbc.co.uk/2/hi/middle_east/3672298.stm (accessed Sept 22, 2004).
15 McCullagh P, Nelder JA. Generalized linear models. London:
and Hall, 1989.
16 Diggle PJ, Heagerty P, Liang KY, Zeger SL. Analysis of longitudinal
data, 2nd edn. New York, NY: Oxford Science
17 Efron B. Bootstrap methods: another look at the jackknife.
Stat 1979; 7: 1–26.
18 Becker SR, Thornton JN, Holder W. Infant and child mortality in
two counties of Liberia: results of a survey in 1988 and trends since 1984. Int J Epidemiol 1993; 22 (suppl 1): S56–63.
19 Taylor WR, Chahnazarian A, Wienman J, et al. Mortality and use
of health services surveys in rural Zaire. Int J Epidemiol 1993; 22 (suppl 1): S15–19.
20 WHO. Health action in crises. http://www.who.int/disasters/stats/baseline.cfm?countryID=62 (accessed Oct 16, 2004).
21 Spiegel PB, Salama P. War and mortality in Kosovo, 1998–99:
an epidemiological testimony. Lancet 2000; 355: 2204–09.
22 CDC. Famine-affected, refugee, and displaced populations:
for public health issues. MMWR Recomm Rep 1992; 41(RR-13): 1–76.
23 CDC. Elevated mortality associated with armed conflict:
Republic of Congo, 2002.
MMWR Morb Mortal Wkly Rep 2003; 52: 469–71.8 www.thelancet.com Published online October 29, 2004
For personal use. Only reproduce with permission from Elsevier Ltd