Mortality data in the age of drones^‡

Abstract

Mortality data plays an essential role in shaping humanitarian, legal and ethical responses to conflict situations. The rise of drone warfare poses new questions regarding the accuracy and reliability of mortality data in conflict. This article examines some of the methodological and political challenges to collecting mortality data in drone warfare, and how the way in which drones are framed in public discourse contributes to these challenges.

Keywords:

• Mortality
• drones
• conflict
• public health
• epidemiology

On the 12 December 2013, a wedding party was travelling to the groom’s village near the town of Rada’a in the al-Bayda’ governorate of Yemen. According to witnesses, doctors and NGOs such as Human Rights Watch and Reprieve, four Hellfire missiles fired by a US drone struck the wedding convoy and killed more than 10 people (Human Rights Watch 2014). The US National Security Council claimed that those killed were members of al-Qa’ida, including a local emir (commander) named Shawqi al-Badani who was the target of the strike – without providing any evidence to substantiate this claim. The families of those who died maintain that they were civilians whose wedding celebration ended in slaughter. Despite compensation paid by the Yemeni Government to the families involved, there has to date been no official US acknowledgement of civilian casualties. The deaths at al-Rada’a raise important questions about collecting mortality data in the age of drones: How is it feasible? What unique challenges does drone warfare pose? How does the framing of drone strikes in public discourse affect whether we believe and how we interpret mortality data?

The importance of mortality data: epidemiology and beyond

Collecting accurate mortality data in conflict situations is vital for planning effective and adaptive humanitarian and public health responses, for pursuing legal remedies in cases of unlawful killing or war crimes, and to meet the ethical imperative to count the human cost of war. Nevertheless, it remains fraught with methodological, logistical and political difficulties (Checchi and Roberts 2008). These are further aggravated by the unique and pernicious nature of drone warfare, which by its nature is secretive, difficult to track and lacks accountability. The justification for US drone warfare is enshrined in secret executive memos written by government lawyers: the contents are not publicly available, but successive administrative officials have maintained that they uphold the legality of lethal drone strikes outside conflict zones.¹ The US Government has also obscured details of casualties from drone strikes (Reprieve 2016), save the occasions when it claims to have killed a specific target; yet even these claims are hard to believe, such as the killing of Qassim al-Raimi, a Yemeni al-Qa’ida leader whom the US has claimed to have killed at least six times (Reprieve 2014). Drone warfare also regularly takes place outside areas of recognized conflict, where there is less international media attention and legal scrutiny: although it has drawn attention in theatres of war such as Afghanistan and Iraq, drone strikes are commonplace in countries such as Pakistan and Somalia where the US is not formally at war (Boyle 2013). This essay will explore why it is important to collect accurate mortality data for drone strikes; the difficulties inherent in this process; and why this will continue to be an issue in the current geopolitical climate and evolving nature of warfare.

Drone strike mortality data is an important subset of overall mortality in conflict situations. The relevance of its legal applications are self-evident, given the uncertainty surrounding the place of drone strikes in international humanitarian and international human rights law (Boyle 2015). It also plays a powerful role in advocacy by anti-drone campaign organizations such as Reprieve, which has brought litigation against the US Government on behalf of the families of drone strike victims (Reprieve 2015). The ethical argument for counting the victims of drone strikes should also be apparent, especially as nations debate the changing nature of warfare and the impact of new technologies on how countries go to war. Yet there is also a public health imperative in collecting this information to assess the unique impact which drone strikes have on mortality and morbidity during conflict. This can inform public health responses, especially when (unlike conventional airstrikes) these strikes may occur outside recognized warzones. Although not a direct measure of morbidity, drone strike deaths may also be a proxy indicator for the potential burden of injury and psychological trauma caused by drone warfare (Alkarama Foundation 2015).

Methodological problems and political obstacles

In conflict scenarios, the collection of mortality data relies on a number of methodological techniques. The epidemiological approach, which uses sampling and surveys to extrapolate the number of deaths and compares this to baseline pre-conflict mortality, stands in contrast to the methodology of organizations such as Iraq Body Count, which only counts deaths where there are cross-checked media reports of a death and its circumstances (Ratnayake et al. 2008). These divergent models can lead to widely differing estimates, such as the well-documented controversy over Iraqi civilian deaths during and after the US invasion: a paper published by public health academics estimated around 650,000 deaths between 2003 and 2006 (around 2.5% of the Iraqi population), whereas the IBC estimated just over 66,000 civilian deaths in the same period (Burnham et al. 2006; Iraq Body Count 2012). Each estimate has been accused of distorting the true figure, which has inevitably played into broader discussions about the legality and morality of the invasion of Iraq; nevertheless, neither figure seemed to trouble the Bush administration, which put the number of dead at 30,000, half the IBC figure (Roberts 2005).

The discrepancies between the two approaches have been dealt with elsewhere (Kaplan 2016), but to an extent they reflect differing priorities concerning mortality data. Both epidemiologists and advocates naturally aim to collect accurate and reliable mortality data. Yet epidemiologists, more concerned with the broader public health impact and without the need to prove individual deaths, rely on their traditional techniques of population sampling. Advocates, on the other hand, may need to adduce their data as evidence; their priority must be complete and substantiated documentation. Documentation of drone strike deaths has so far taken mainly the latter approach: anti-drone advocacy focuses (with obvious justification) on the legality of drone warfare and civilian deaths, hence the need for precise identification of the victims and verification of their civilian status. Furthermore, the flaws in using an epidemiological approach to count drone deaths are apparent. Given the relative infrequency of drone deaths, a population sampling methodology might not capture any deaths at all. It is also often unclear to witnesses on the ground whether a strike came from a drone or a conventional warplane, rendering the cause of death ambiguous. Finally, most drone warfare takes place where epidemiological surveys are inherently difficult and dangerous, such as Yemen, Pakistan and Somalia.

Tracking drone deaths consequently depends on collecting media reports of strikes, a task which has been undertaken by organizations such as the Bureau of Investigative Journalism and Long War Journal. It is not an easy endeavour. As a report by Columbia Law School indicates, there are a number of flaws inherent in media surveillance: not all drone strikes are covered consistently, there is often a language barrier between sources and those collecting them, reports of casualties are often provided by anonymous sources, and so on (Grut et al. 2012). The same report analysed media reports of drone casualties in Pakistan in 2011 and arrived at very different figures to those obtained by the three main organizations tracking drone deaths using this methodology. Well-documented drone deaths such as the wedding party strike in Rada’a are the exception: more commonly, a report will run something along the lines of this CNN news item from August 2013, which cites anonymous security officials and vague ‘links’ to al Qa’ida and does not account for two of those killed:

In central Yemen’s Mareb province, eight people were killed in an early morning drone strike, including four with links to al Qaeda, local security officials said. Two civilians were among those killed in the strike, which targeted two vehicles, the officials said. (Almasmari et al. 2013)

The corollary to these methodological difficulties is the political intransigence which further hampers efforts to collect accurate and reliable drone mortality data. The US has historically been reluctant to release any casualty data relating to the drone programme. In July 2016, the Obama administration released its first official figures, claiming that between 64 and 116 non-combatants had been killed between January 2009 and the end of 2015 in drone strikes in Pakistan, Yemen, Somalia and Libya (Ackerman 2016). These numbers were widely criticized by organizations such as the Bureau of Investigative Journalism, which cited its own figures of 380 to 801 civilian casualties for the same time period (Serle 2016). The key issue here is that as long as the US Government refuses to be transparent about how it classifies and monitors civilian casualties, there is no reliable way to compare its figures to those of tracking organizations such as the Bureau, with deleterious consequences for the credibility of its casualty estimates. The governments of countries where drone strikes take place, such as Pakistan and Yemen, can sometimes also display a reluctance to quantify drone mortality on their own soil, and have faced criticism from human rights groups for underplaying the extent of civilian mortality (BBC News 2013a). This is often due to sensitivity around the implications of US drone strikes for territorial sovereignty: governments may publicly condemn US strikes while tacitly agreeing to them, and as a consequence might avoid emphasis on mortality data which could draw unwelcome scrutiny to their own permissiveness (BBC News 2013b).

Barriers to belief: how drones are framed in public discourse

Beyond the difficulty of collecting accurate mortality data, the question remains what we as a society are able or prepared to believe about drone strikes. As Oliver Kearns writes, the existence of drone strikes in public discourse is characterized by ‘absences, of documentation of the violence and violated bodies’. (Kearns 2017). The practice of drone warfare and its consequences are concealed from public view by three factors: the official secrecy surrounding drone policy and casualty figures; the geographical remoteness of strikes and the frequent absence of mainstream news media to report on them; and the language used by governments to talk about drone strikes, which construes them as ‘surgical’, ‘precise’ and ‘targeted’.

While official secrecy obscures the release of mortality data, the geographical and semantic remoteness of drone strikes renders such data difficult to believe or make sense of. Kearns discusses the stereotypical image of the burned-out car often used to illustrate media stories about drone warfare: unlike conventional warfare, where pictures of casualties are a common feature, reporting on drone deaths is more often defined by the absence of a body. The unusual theatres of drone warfare exacerbate this cognitive absence surrounding drone mortality: the public is accustomed to processing conflict mortality data in countries such as Iraq and Afghanistan, but perhaps not Pakistan, a nominal US ally and partner in the War on Terror, where between 424 and 969 civilians have been killed by US drone strikes in the years 2004 to 2017, according to the Bureau of Investigative Journalism (Bureau of Investigative Journalism 2018). Where drone strikes are alluded to by government officials, the euphemisms which are used serve to distance them from violence and death. As Elke Schwarz describes, drones are ‘repeatedly … referred to as instruments that enable cancerous terrorist cells to be eliminated with surgical precision’. (Schwarz 2015). The body has been relocated from the dead body of the drone victim to the sick body of the state which requires the curative attentions of the drone operator. All this serves to distance us from drone deaths, making mortality data even harder to believe.

Conclusion

In September 2017, reports suggested that President Trump was seeking to change the Obama-era policy underpinning constraints on drone strikes (Chesney 2017). Although – so far – no changes have been made, the prospect of any further expansion of the US drone programme is alarming. It highlights the need for improved mortality data collection in an age when drone warfare is becoming increasingly prevalent. This data collection is crucial for a public health assessment of the impact of new warfare technologies, and for any potential legal challenges, yet mortality data is rendered ‘unbelievable’ by methodological difficulties, political obstacles and the way in which drone deaths are framed in public discourse. While the efforts of organizations such as the Bureau of Investigative Journalism are admirable, the solution inevitably lies in transparent and prompt reporting of civilian casualties by the governments who are responsible for them. In the shadows of a secret war, where dead wedding guests lie uncounted, the time has come for a full reckoning.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes on contributor

Elspeth Carruthers is a third-year medical student at St George’s, University of London. She previously worked for Reprieve, a legal charity which represents victims of drone strikes and their families.

Notes

1. In May 2013 the Obama administration issued the Presidential Policy Guidance on drone warfare, also known as the drone ‘playbook’, which set certain restrictions on strikes conducted outside ‘areas of active hostilities’. However, this was a policy document rather than a legal standard.