Intended for healthcare professionals


Circumstances around weapon injury in Cambodia after departure of a peacekeeping force: prospective cohort study

BMJ 1999; 319 doi: (Published 14 August 1999) Cite this as: BMJ 1999;319:412
  1. David R Meddings, epidemiologist (dmeddings{at},
  2. Stephanie M O'Connor, head nurseb
  1. a Unit of the Chief Medical Officer, International Committee of the Red Cross, 19 avenue de la Paix, 1202 Geneva, Switzerland
  2. b Health Operations Division, International Committee of the Red Cross, 19 avenue de la Paix, 1202 Geneva
  1. Correspondence to: Dr Meddings
  • Accepted 30 June 1999


Objective: To examine the circumstances surrounding weapon injury and combatant status of those injured by weapons.

Design: Prospective cohort study.

Setting: Northwestern Cambodia after departure of United Nations peacekeeping force.

Subjects: 863 people admitted to hospital for weapon injuries over 12 months.

Main outcome measures: Annual incidence of weapon injury by time period; proportions of injuries inflicted as a result of interfactional combat (combat injuries) and outside such combat (non-combat injuries) by combatant status and weapon type.

Results: The annual incidence of weapon injuries was higher than the rate observed before the peacekeeping operation. 30% of weapon injuries occurred in contexts other than interfactional combat. Most commonly these were firearm injuries inflicted intentionally on civilians. Civilians accounted for 71% of those with non-combat injuries, 42% of those with combat related injuries, and 51% of those with weapon injuries of either type.

Conclusions: The incidence of weapon injuries remained high when the disarmament component of a peacekeeping operation achieved only limited success. Furthermore, injuries occurring outside the context of interfactional combat accounted for a substantial proportion of all weapon injuries, were experienced disproportionately by civilians, and were most likely to entail the intentional use of a firearm against a civilian.

Key messages

  • The study took place in Cambodia after a United Nations peacekeeping operation that achieved only limited success in disarmament

  • A substantial proportion of weapon injuries was inflicted in contexts unrelated to interfactional combat

  • These injuries were most commonly firearm injuries inflicted intentionally on civilians

  • Widespread availability of weapons can facilitate social violence

Editorial by Smith


In many areas of the world military weapons are widely available.1 2 This has been argued to contribute to regional instability, increased civilian injuries, and violence that is not directly related to interfactional combat.3 4

In 1990 the International Committee of the Red Cross began supporting Mongkol Borei hospital in Banteay Meanchey province in northwestern Cambodia. The hospital was the only facility in the region with the capacity to provide surgical care to people injured by weapons. The peace accords of 1991 led to the arrival of the United Nations Transitional Authority in Cambodia in March 1992. By October 1992 this international peacekeeping operation was supposed to have disarmed and demobilised 70% of Cambodia's combatant factions.5 In November 1992 it announced that it could not meet this disarmament target because of non-compliance of some troops. Around 25-50% of the troops are believed to have been disarmed.5

We prospectively examined combatant status and the circumstances surrounding injury for people with weapons injuries admitted to Mongkol Borei hospital during a 12 month period after the departure of the United Nations Transitional Authority in Cambodia.

Table 1.

Characteristics of patients and type of injury. Values are numbers (percentages) unless stated otherwise

View this table:

Patients and methods

Between 1 March 1994, five months after the departure of the United Nations Transitional Authority in Cambodia, and 28 February 1995, all people admitted to Mongkol Borei hospital with weapon injuries received a structured interview eliciting demographic information, combatant status, and circumstances surrounding injury.

We categorised injuries as combat and non-combat on the basis of the context of injury and as occurring in civilians or military staff on the basis of combatant status. Combat injuries were defined as those sustained during interfactional combat or from stepping on a landmine. Injuries from all other contexts were classed as non-combat injuries, and subcategorised into those resulting from interpersonal violence or by accident. On the rare occasions that the classification of injury was ambiguous the category was assigned by one of us (SMO).

Incidence of weapon injury was calculated using admissions data from January 1991 to February 1995 and figures for the population of Banteay Meanchey obtained from the United Nations High Commissioner for Refugees. These rates were calculated for our study period and the periods preceding and during the mandate of the United Nations Transitional Authority in Cambodia.


Monthly admissions for weapon injury to Mongkol Borei hospital, 1991-5


Table 1 shows the characteristics of the 863 people injured during our study. Mine injuries were most common, followed by injuries due to fragmenting munitions (mortars, bombs, or grenades) and firearms.

The figure shows the monthly admissions for weapon injury to Mongkol Borei hospital from January 1991 to February 1995. Injury rates varied seasonally. They were comparable before the arrival and after the departure of the United Nations Transitional Authority in Cambodia and reduced during its presence.

Table 2 shows mean annual and seasonally adjusted mean annual incidence of weapon injury for the study period and before and during the mandate of the United Nations Transitional Authority in Cambodia. Both rates were higher during the study period than before the arrival of the United Nations.

Table 2.

Mean annual and seasonally adjusted mean annual incidence of weapon injury in Mongkol Borei hospital by period

View this table:

Table 3 presents our data disaggregated by injury context, combatant status, and weapon type. The box summarises the important points.


Nearly one in three people sustained non-combat injuries. Moreover, intentional firearm injuries among civilians was by far the largest category of non-combat injury. This supports the contention that widespread weapon availability, in the context of a protracted conflict, is associated with a high rate of social violence

Limitations and potential biases

Several limitations should be considered. We inferred that weapon availability was high without measuring it. However, a United Nations military survey in December 1991 reported that the Cambodian combatant factions possessed over 320 000 weapons and over 80 million rounds of ammunition.5

Injuries were underascertained since some people die before reaching care or survive without presenting for care.6 This underascertainment may have been affected by injury context or combatant status. If so, it is difficult to know in which direction such factors would operate. However, we think that access to care of military staff was probably no worse, and probably better, than that of civilians. Therefore, our findings may have been biased towards underestimating civilian casualties and non-combat injuries.

Table 3.

Number of weapon injuries by context, combatant status, and weapon type

View this table:

Important points from table 3

All injuries
  • 59% of people injured were civilians or did not sustain their injuries as a direct result of active fighting between armed factions, or both

  • 51% of people injured were civilians

  • 37% of all those injured were injured by mines

Non-combat injuries
  • 30% of people sustained their injuries as a result of something other than active fighting between armed factions

  • 71% of all people with non-combat injuries were civilians

  • 67% of people with non-combat injuries from firearms were civilians injured as a result of interpersonal violence

  • 79% of non-combat injuries to military staff resulted from accidents

Combat injuries
  • 46% of injuries related to combat were from mines

  • 42% of all people with combat injuries were civilians

  • 78% of civilians with combat injuries which required an act of volition—that is, from all weapons except landmines—were injured by fragmenting munitions (bombs, shells, and grenades)

  • 60% of military staff with such injuries were injured by firearms and 38% by fragmenting munitions

Some injuries may have been misclassified. However, this was unlikely to be an important source of error. Interviews were conducted by one of two trained Cambodian colleagues who were debriefed daily, allowing ambiguities to be resolved while the patient was still in hospital Finally, because classification criteria were unambiguous, few cases required designation of category.

Mine injuries

Our results add to other evidence of the extent of the problem of antipersonnel mines in Cambodia.7 8 Such data were instrumental in leading to the treaty on antipersonnel mines signed by 124 countries in Ottawa in December 1997.

Civilian casualties

The weapon type causing most civilian injuries was not mines but fragmenting munitions followed by firearms, weapons requiring an act of volition on the part of the user. Some authors argue that civilian casualties constitute a pressing humanitarian issue 9 10 and advocate describing the epidemiology of the issue to address it with a public health approach.11 12

The likelihood of civilian injury depends on the context in which a given injury was inflicted. These different contexts have very different implications for addressing civilian casualties. Fragmenting munitions accounted for nearly 80% of civilian casualties resulting from intentional weapon use in the context of combat. Some preventive strategies have been laid out in a recent review.11 Combatants also require better training in use of these weapons and in the rights of civilians to protection under international humanitarian law. A consideration in these efforts should be that combatants might feel less responsible for civilian casualties provided a distance separates them from the victims of their weapons.13 14

Preventing civilian casualties in non-combat contexts entails different considerations. Firearms accounted for most of these injuries, and almost 90% of civilian casualties inflicted with firearms occurred during interpersonal violence. Over a third of civilian casualties in non-combat contexts occurred by accident, and most of these concerned fragmenting munitions, namely, curious children pulling pins from hand grenades.

Despite different mechanisms of injury, we believe that these results provide support for the contention that reducing weapon availability is essential to curtailing social violence and providing conditions requisite for social development.2 4 Despite an estimated cost of $1.5bn, the disarmament component of the mandate of the United Nations Transitional Authority in Cambodia was abandoned.5 Our findings reflect the human cost of this decision. It is a cost we believe should be considered in international policies concerned with arms availability and transfer.


Contributors: DRM conducted the analysis and interpretation of the data, discussed core ideas, and wrote the paper. SMO initiated and coordinated the collection of data, discussed core ideas and interpretation of the data, and participated in writing the paper. DRM is guarantor of the paper


  • Funding No additional funding.

  • 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.
  12. 12.
  13. 13.
  14. 14.