Intended for healthcare professionals


Counting the dead in Iraq

BMJ 2005; 330 doi: (Published 10 March 2005) Cite this as: BMJ 2005;330:550
  1. Klim McPherson (Klim.McPherson{at}, visiting professor of public health epidemiology
  1. Department of Obstetrics and Gynaecology, Oxford University, Oxford OX1 2JD

We need to know how many people have died

Counting the dead is intrinsic to civilised society. Understanding the causes of death is a core public health responsibility. The government's white paper on public health emphasises the vital role of assessing the impact on health of all public policy.1 This is well recognised, and yet neither the public nor public health professionals are able to obtain reliable and officially endorsed information about the extent of civilian deaths attributable to the allied invasion of Iraq. Estimates vary between tens and hundreds of thousands.

These estimates come from reports in the press, or counting bodies admitted to hospitals, ( as well as surveys. The former are likely to be inaccurate and to underestimate the true numbers and do not easily allow for reliable attribution between, for example, violent and natural causes. Public access to reliable data on mortality is important. The policy being assessed—the allied invasion of Iraq—was justified largely on grounds of democratic supremacy. Voters in the countries that initiated the war, and others—not least in Iraq itself—are denied a reliable evaluation of a key indicator of the success of that policy. This is unacceptable.

Instead the UK government's policy was first not to count at all, and then to rely publicly on extremely limited data available from the Iraqi Ministry of Health. This follows US government policy; famously encapsulated by General Tommy Franks of the US Central Command “We don't do body counts.”2 Its inadequacy was emphasised after the publication of a representative household survey that estimated 100 000 excess deaths since the 2003 invasion.3 The government rejected this survey and its estimates as unreliable; in part absurdly because statistical extrapolation from samples was thought invalid.4 Imprecise they are, but to a known extent. These are unique estimates from a dispassionate survey conducted in the most dangerous of epidemiological conditions. Hence the estimates, as far as they can go, are unlikely to be biased, even allowing for the reinstatement of Falluja. To confuse imprecision with bias is unjustified.

The methods for counting the dead in such circumstances are well established and cannot rely on incidental reports or assessments in hospital mortuaries alone. They require first hand verbal autopsies,5 which should be reliably obtained so that extrapolation to the population is possible, as Roberts et al had done. They also require some linkage with unclassified data on military offensives.6 Although active surveillance of this kind is extremely difficult in the context of such violence, even limited household surveys are essential so long as they are systematic. Such data can then be combined with information from passive sources to establish a more accurate overall assessment.

Counting casualties accurately can help to save lives both currently and in the future. Understanding the burden of death, injury, disease, and trauma that the population is currently suffering enables proper planning of war, and health, and in assessing local responses appropriately. In the future this should help government and military planners to assess the likely humanitarian implications of conflict.

The plain fact is that an estimate of 100 000 excess deaths attributable to the invasion of Iraq is alarming. This is already half the death toll of Hiroshima.7 Apart from the practical arguments, the principled ones stand and will always stand. Have we not learnt any lessons from the history of sweeping alarming numbers of deaths under the carpet? This is not something about which there can be any political discretion 60 years after Auschwitz. The UK government, acting on our behalf, ought to offer reasoned criticism of the existing estimates. It should pursue their public health responsibilities to count the casualties by using modern methods. Democracy requires this, as does proper responsibility under the Geneva Conventions.

The sources the government prefers are likely to be seriously biased for several reasons. They do not take into account deaths during the first 12 months since the invasion8; only violence related deaths reported through the health system are taken into account (very likely to lead to an underestimate9); and non-violent deaths due to the destruction of war are not taken into account. Furthermore, even these limited figures are no longer being released on request.10

Apparently the defence ministry has set up an appraisal group,11 but we urgently await transparency and public accountability. The time elapsed since the announcement is already longer than it took to conduct the field survey last year. Electorates, in Iraq and elsewhere, have a right to know. To procrastinate further for no good reason is to devalue public health processes, not to mention Iraqi lives. As public health professionals we need to know the health costs.


  • See also News p 557

  • Competing interests None declared.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
View Abstract

Log in

Log in through your institution


* For online subscription