Bilukha OO, Brennan M, Woodruff BA. Death and Injury From Landmines and Unexploded Ordnance in Afghanistan. JAMA. 2003;290(5):650–653. doi:10.1001/jama.290.5.650
Author Affiliations: Epidemic Intelligence Service, Epidemiology Program Office (Dr Bilukha) and International Emergency and Refugee Health Branch, Division of Emergency and Environmental Health Services, National Center for Environmental Health (Drs Bilukha, Brennan, and Woodruff), Centers for Disease Control and Prevention, Atlanta, Ga.
Context Afghanistan is one of the countries most affected by injuries due to
landmines and unexploded ordnance.
Objective To understand the epidemiological patterns and risk factors for injury
due to landmines and unexploded ordnance.
Design and Setting Analysis of surveillance data on landmine and unexploded ordnance injuries
in Afghanistan collected by the International Committee of the Red Cross in
390 health facilities in Afghanistan. Surveillance data were used to describe
injury trends, injury types, demographics, and risk behaviors of those injured
and explosive types related to landmine and unexploded ordnance incidents.
Participants A total of 1636 individuals injured by landmines and unexploded ordnance,
March 2001 through June 2002.
Results Eighty-one percent of those injured were civilians, 91.6% were men and
boys, and 45.9% were younger than 16 years. Children were more likely to be
injured by unexploded ordnance (which includes grenades, bombs, mortar shells,
and cluster munitions), whereas adults were injured mostly by landmines. The
most common risk behaviors for children were playing and tending animals;
for adults, these risk behaviors were military activity and activities of
economic necessity (eg, farming, traveling). The case-fatality rate of 9.4%
is probably underestimated because surveillance predominantly detects those
who survive long enough to receive medical care.
Conclusions Landmine risk education should focus on hazards due to unexploded ordnance
for children and on landmine hazards for adults and should address age-specific
risk behaviors. Expanding community-based and clinic-based reporting will
improve the sensitivity and representativeness of surveillance.
Landmines and unexploded ordnance pose a significant public health risk
and economic threat worldwide.1,2 Approximately
60 million to 70 million landmines are placed in about 70 countries,3 and an estimated 24 000 individuals, mostly civilians,
are killed or injured by landmines and unexploded ordnance worldwide every
year.4 Unexploded ordnance includes military
explosive munitions such as grenades, bombs, mortar shells and cluster munitions,
which have been deployed or scattered during military activities but have
failed to detonate.
During 2000 and 2001, Afghanistan had the most reported landmine and
unexploded ordnance casualties in the world.5 Most
of the 7 million to 8 million landmines in Afghanistan were laid during the
Soviet occupation between 1978 and 1989 and the subsequent civil war.3 The coalition air strikes in the fall of 2001 exacerbated
the problem by deploying a new type of ordnance—cluster bomblets.5,6 Because the bomblets were bright yellow,
concerns existed that children might mistake them for the humanitarian daily
rations airdropped during the early stages of the conflict. Another feature
of recent conflicts in Afghanistan and Iraq has been the aerial bombardment
of munitions dumps. This resulted in scattering of explosives over wide areas.
In such situations, injuries due to unexploded ordnance may be as much of
a public health threat as those due to landmines.
We undertook this study to evaluate landmine and unexploded ordnance
injury surveillance in Afghanistan, to document the magnitude of these injuries
and deaths between March 2001 and June 2002, and to describe epidemiological
patterns of and risk factors for injury.
We obtained data on landmine and unexploded ordnance casualties from
March 2001 through June 2002 from the International Committee of the Red Cross.
Duplicate entries were excluded, and statistical analyses were performed using
JMP software (Release 5.0, SAS Institute Inc, Cary, NC).
These data were obtained from a clinic-based surveillance system operated
by the International Committee of the Red Cross. There are an estimated 905
functioning health facilities in Afghanistan.7 Among
these 905, the International Committee of the Red Cross identified those facilities
most likely to see injuries caused by landmines and unexploded ordnance. These
include major referral hospitals, provincial and district hospitals, specialized
surgical or rehabilitation clinics, facilities supported by nongovernmental
organizations, and basic health centers in areas contaminated by landmines
or unexploded ordnance. The current surveillance system includes 390 clinics
and hospitals throughout the country8 and is
believed to include most facilities likely to see injuries due to landmines
and unexploded ordnance. Not included in the system are specialized facilities,
such as tuberculosis or malaria treatment centers, facilities located in areas
with no known landmine or unexploded ordnance problems, facilities in which
access is difficult or the situation is insecure, and facilities not wishing
to participate in the system.
Surveillance uses a standard data collection form compatible with the
Information Management System for Mine Action, recommended and widely used
by the United Nations Mine Action Service worldwide.5 Those
sustaining injuries from landmines and unexploded ordnance, or their family
members, are interviewed by trained staff at the health facility, and the
forms are forwarded to International Committee of the Red Cross headquarters
in Kabul each month.
As participating health facilities were considered neither a random
nor representative sample of all health facilities in Afghanistan, the injuries
detected are likewise not considered representative of all relevant injuries.
Nonetheless, for participating facilities, reported injuries are considered
a complete list of all injuries due to landmines and unexploded ordnance seen
in those facilities. For these reasons, measures of precision, such as confidence
intervals, were not calculated nor was hypothesis testing performed to test
the statistical significance of differences between subgroups. Therefore only
substantial observed differences between subgroups are highlighted in this
This analysis is based on data on 1636 individuals injured by landmines
and unexploded ordnance. An average of 102 new injuries were reported each
month; the lowest number was reported in October 2001 (Figure 1). A pronounced increase in injuries from unexploded cluster
munitions began in October 2001, when the conflict began between the Taliban
government and coalition forces. Cluster munitions are weaponized containers
that are intended to break open in mid air and disperse smaller munitions
or submunitions that are intened to explode just before or at impact.
Forty-six percent of those injured were younger than 16 years. The highest
number of injuries was seen among those aged 7 to 15 years. Reported injuries
in all age groups were mostly among males (91.6%) and civilians (81.2%) (Table 1).
Children were more likely to be injured by unexploded ordnance; injuries
among adults were caused predominantly by landmines. The case-fatality rate
was 9.4% and varied little across the age groups (Table 1).
Among injuries to children and adolescents younger than 16 years for
whom activity was known, playing and tending animals were the most common
activities, accounting for 49% of injuries. Military activity, traveling,
and farming were the most common risk behaviors among persons 16 years or
older. About 11% of those injured, mostly adults, reported that they knowingly
took the risk of going into a dangerous area. About 11% of injured persons
reported having received some landmine risk education before the injury (Table 1).
A stratified analysis assessed the relationship of age and activity
to explosive type. For each of the activities of economic necessity (ie, farming;
tending animals; traveling; collecting wood, food, or water), children were
more likely to be injured by unexploded ordnance and adults were more likely
to be injured by landmines. On the other hand, both children and adults injured
while playing or tampering with explosives were injured predominantly by unexploded
ordnance (Table 2).
The results of this study demonstrate that landmines and unexploded
ordnance remain a serious public health threat in Afghanistan. As has been
seen elsewhere during postconflict periods,9- 11 most
of the injured are civilians, with children and adolescents being at highest
risk. In addition, surveillance data presented in this report concern only
acute injuries and do not address long-term physical disability and mental
health concerns, which may add substantially to both the economic and public
Furthermore, as our data show, because landmines and unexploded ordnance
seriously hinder the simple activities of economic necessity, they may substantially
undermine postconflict recovery. The most important risk factors in adults
(besides military activity) were activities of economic necessity: farming,
tending animals, traveling, and collecting food, wood, and water. However,
this risk is not limited to adults. Afghan children begin to be involved in
herding at a relatively young age; this is reflected in our data, which show
that 13% of injured children aged 6 years or younger and one quarter of injured
children aged 7 to 15 years were injured while involved in this activity.
These children may be at particularly high risk if herding prevents them from
attending school because many landmine risk education programs targeted toward
children are school-based. Special programs focusing on children engaged in
herding and not at school may be necessary.
Although prior reports and world opinion often pay greater attention
to landmines, our data show that more than half of all injuries and deaths
in Afghanistan were caused by unexploded ordnance rather than by landmines.
Injuries from unexploded ordnance were more prevalent in children and adolescents
younger than 16 years, regardless of risk behavior. Because unexploded ordnance
are more visible than landmines, unexploded ordnance–related injuries
are preventable and should become a priority in landmine risk education efforts.
The proportion of persons who reported knowing that the area was dangerous,
that the area was marked for landmines and unexploded ordnance, or that they
had received mine risk education was low. However, using these results to
evaluate mine risk education is difficult because no estimates exist of the
coverage of mine risk education in the general population to permit calculation
of injury rates among persons who did and did not receive such education.
A rigorous evaluation of the effectiveness of landmine risk education in reducing
injuries due to landmines and unexploded ordnance is needed in Afghanistan
and other countries.
The surveillance system detected an increase in injuries from cluster
munitions during October 2001, which coincided with the start of bombings
by coalition forces. This increase gradually subsided after February 2002.
These data are consistent with reports about increased danger to civilians
from unexploded cluster bomblets.6 The decrease
in such injuries may be attributed in part to rapid clearance response measures
undertaken by the Mine Action Center for Afghanistan after receiving information
from the US military about the location of coalition bomb strikes.5
The results of this study are subject to several important limitations.
The actual numbers of those injured and killed by landmines and unexploded
ordnance in Afghanistan are likely to be substantially higher than that reported
by surveillance data because the clinic-based surveillance system is likely
to miss those who die before reaching a clinic, those whose injuries are too
minor to seek medical care, and those living in areas with little or no access
to health facilities that are involved in the surveillance system. Because
many fatal injuries are probably missed, a case-fatality rate of 9.4% derived
from surveillance data is likely to be substantially underestimated. Previous
studies using population survey methods found a case-fatality rate of 55%
in Afghanistan,12 41% in Bosnia,12 42%
and 48% in Mozambique,12,13 and
31% in Cambodia.12
Time trends in injuries seen in the surveillance data should be interpreted
with caution because of low sensitivity and differential system coverage over
time. For example, the low number of injuries recorded during September and
October 2001 may be due in part to disruption of surveillance activity before
and during the coalition strikes.
The disproportionate number of injuries among men may be because women
in Afghanistan are more restricted than men in their mobility and thus are
less likely to engage in activities that put them at risk for landmine or
unexploded ordnance injuries. In addition, some of this observed sex disparity
may result from decreased likelihood of injured females to receive medical
care or to be interviewed and recorded by the surveillance system.
The absence of reliable data on the age distribution of the Afghan population
did not allow for calculating age-specific injury rates. However, by conservative
estimates it is unlikely that the group of 7- to 15-year-olds accounts for
more than 25% of total population. Therefore, the finding that 40% of all
injuries occurred in this age group suggests that this group may indeed be
at higher risk.
Expanding clinic-based surveillance to include nonparticipating health
facilities would allow recording of more injuries, thus further increasing
sensitivity. Using alternative methods for data collection, such as community-based
surveillance or population-based surveys would allow for recording more representative
data, including fatal and minor injuries, and injuries among women and girls.
Surveillance data suggest that landmine risk education should focus
on landmine avoidance among adults and avoidance of unexploded ordnance among
children and adolescents. Surveillance data from Iraq and other countries
where munition dumps have been bombed may show a similar need to focus mine
risk education programs on the risks of unexploded ordnance injury as much
as landmine injury, particularly among children.