Death and Injury from Landmines and Unexploded Ordnance in Afghanistan

BRIEF REPORT

Death and Injury From Landmines and Unexploded Ordnance in Afghanistan

Oleg O. Bilukha, MD, PhD Context Afghanistan is one of the countries most affected by injuries due to land- Muireann Brennan, MD, MPH mines and unexploded ordnance. Bradley A. Woodruff, MD, MPH Objective To understand the epidemiological patterns and risk factors for injury due to landmines and unexploded ordnance. ANDMINES AND UNEXPLODED ORD- Design and Setting Analysis of surveillance data on landmine and unexploded ordnance pose a significant public nance injuries in Afghanistan collected by the International Committee of the Red Cross health risk and economic threat in 390 health facilities in Afghanistan. Surveillance data were used to describe injury worldwide.1,2 Approximately 60 trends, injury types, demographics, and risk behaviors of those injured and explosive millionL to 70 million landmines are types related to landmine and unexploded ordnance incidents. placed in about 70 countries,3 and an esti- Participants A total of 1636 individuals injured by landmines and unexploded ord- mated 24000 individuals, mostly civil- nance, March 2001 through June 2002. ians, are killed or injured by landmines Results Eighty-one percent of those injured were civilians, 91.6% were men and boys, and unexploded ordnance worldwide and 45.9% were younger than 16 years. Children were more likely to be injured by every year.4 Unexploded ordnance unexploded ordnance (which includes grenades, bombs, mortar shells, and cluster mu- includes military explosive munitions nitions), whereas adults were injured mostly by landmines. The most common risk be- such as grenades, bombs, mortar shells haviors for children were playing and tending animals; for adults, these risk behaviors and cluster munitions, which have been were military activity and activities of economic necessity (eg, farming, traveling). The deployed or scattered during military case-fatality rate of 9.4% is probably underestimated because surveillance predominantly detects those who survive long enough to receive medical care. activities but have failed to detonate. During 2000 and 2001, Afghanistan Conclusions Landmine risk education should focus on hazards due to unexploded had the most reported landmine and un- ordnance for children and on landmine hazards for adults and should address age- specific risk behaviors. Expanding community-based and clinic-based reporting will exploded ordnance casualties in the improve the sensitivity and representativeness of surveillance. world.5 Most of the 7 million to 8 mil- JAMA. 2003;290:650-653 www.jama.com lion landmines in Afghanistan were laid during the Soviet occupation between 1978 and 1989 and the subsequent civil nance injury surveillance in Afghani- the Red Cross. There are an estimated war.3 The coalition air strikes in the fall stan, to document the magnitude of 905 functioning health facilities in Af- of 2001 exacerbated the problem by these injuries and deaths between ghanistan.7 Among these 905, the In- deploying a new type of ordnance— March 2001 and June 2002, and to de- ternational Committee of the Red Cross cluster bomblets.5,6 Because the bomb- scribe epidemiological patterns of and identified those facilities most likely to lets were bright yellow, concerns ex- risk factors for injury. see injuries caused by landmines and isted that children might mistake them unexploded ordnance. These include for the humanitarian daily rations air- METHODS major referral hospitals, provincial and dropped during the early stages of the We obtained data on landmine and un- district hospitals, specialized surgical conflict. Another feature of recent con- exploded ordnance casualties from

flicts in Afghanistan and Iraq has been March 2001 through June 2002 from Author Affiliations: Epidemic Intelligence Service, Epi- the aerial bombardment of munitions the International Committee of the demiology Program Office (Dr Bilukha) and Interna- dumps. This resulted in scattering of ex- Red Cross. Duplicate entries were ex- tional Emergency and Refugee Health Branch, Divi- sion of Emergency and Environmental Health Services, plosives over wide areas. In such situa- cluded, and statistical analyses were National Center for Environmental Health (Drs Bi- tions, injuries due to unexploded ord- performed using JMP software (Re- lukha, Brennan, and Woodruff ), Centers for Disease Control and Prevention, Atlanta, Ga. nance may be as much of a public health lease 5.0, SAS Institute Inc, Cary, NC). Corresponding Author and Reprints: Oleg O. Bi- threat as those due to landmines. These data were obtained from a lukha, MD, PhD, National Center for Environmental Health, Centers for Disease Control and Prevention, We undertook this study to evalu- clinic-based surveillance system oper- 4770 Buford Hwy, Mailstop F-48, Atlanta, GA 30341 ate landmine and unexploded ord- ated by the International Committee of (e-mail: [email protected]).

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LANDMINES AND UNEXPLODED ORDNANCE

or rehabilitation clinics, facilities sup- Figure. Individuals With Injuries From Landmines and Unexploded Ordnance per Month, by ported by nongovernmental organiza- Explosive Type, in Afghanistan, March 2001-June 2002 tions, and basic health centers in areas contaminated by landmines or unex- Explosive Type ploded ordnance. The current surveil- Other/Unknown Landmine∗ Unexploded Ordnance† Other Unexploded Cluster Munition‡ lance system includes 390 clinics and Than Landmines and Cluster Munition 8 hospitals throughout the country and 200

is believed to include most facilities 180

likely to see injuries due to landmines 160

and unexploded ordnance. Not in- 140 cluded in the system are specialized fa- 120 cilities, such as tuberculosis or ma- 100 laria treatment centers, facilities located in areas with no known landmine or un- 80 exploded ordnance problems, facili- No. of Injured 60 ties in which access is difficult or the 40 situation is insecure, and facilities not 20 0 wishing to participate in the system. Mar May Jul Sep Nov Jan Mar May Jul Surveillance uses a standard data col- 2001 2002 lection form compatible with the In- formation Management System for Asterisk indicates landmine includes antipersonnel and antitank landmines. Dagger indicates munitions such as grenades, bombs, or mortar shells. Double dagger indicates weaponized containers that are intended to Mine Action, recommended and widely break open in mid air and disperse smaller munitions or submunitions. used by the United Nations Mine Ac- tion Service worldwide.5 Those sus- taining injuries from landmines and un- month; the lowest number was re- farming were the most common risk be- exploded ordnance, or their family ported in October 2001 (FIGURE). A haviors among persons 16 years or members, are interviewed by trained pronounced increase in injuries from older. About 11% of those injured, staff at the health facility, and the forms unexploded cluster munitions began in mostly adults, reported that they know- are forwarded to International Com- October 2001, when the conflict be- ingly took the risk of going into a dan- mittee of the Red Cross headquarters gan between the Taliban government gerous area. About 11% of injured per- in Kabul each month. and coalition forces. Cluster muni- sons reported having received some As participating health facilities were tions are weaponized containers that are landmine risk education before the in- considered neither a random nor rep- intended to break open in mid air and jury (Table 1). resentative sample of all health facili- disperse smaller munitions or submu- A stratified analysis assessed the re- ties in Afghanistan, the injuries detected nitions that are intened to explode just lationship of age and activity to explo- are likewise not considered representa- before or at impact. sive type. For each of the activities of eco- tive of all relevant injuries. Nonethe- Forty-six percent of those injured nomic necessity (ie, farming; tending less, for participating facilities, reported were younger than 16 years. The high- animals; traveling; collecting wood, food, injuries are considered a complete list est number of injuries was seen among or water), children were more likely to of all injuries due to landmines and unex- those aged 7 to 15 years. Reported in- be injured by unexploded ordnance and ploded ordnance seen in those facili- juries in all age groups were mostly adults were more likely to be injured by ties. For these reasons, measures of pre- among males (91.6%) and civilians landmines. On the other hand, both chil- cision, such as confidence intervals, were (81.2%) (TABLE 1). dren and adults injured while playing or not calculated nor was hypothesis test- Children were more likely to be tampering with explosives were in- ing performed to test the statistical sig- injured by unexploded ordnance; jured predominantly by unexploded nificance of differences between sub- injuries among adults were caused ordnance (TABLE 2). groups. Therefore only substantial predominantly by landmines. The case- observed differences between sub- fatality rate was 9.4% and varied little COMMENT groups are highlighted in this article. across the age groups (Table 1). The results of this study demonstrate Among injuries to children and ado- that landmines and unexploded ord- RESULTS lescents younger than 16 years for nance remain a serious public health This analysis is based on data on 1636 whom activity was known, playing and threat in Afghanistan. As has been seen individuals injured by landmines and tending animals were the most com- elsewhere during postconflict peri- unexploded ordnance. An average of mon activities, accounting for 49% of ods,9-11 most of the injured are civil- 102 new injuries were reported each injuries. Military activity, traveling, and ians, with children and adolescents

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LANDMINES AND UNEXPLODED ORDNANCE

ploded ordnance rather than by land- Table 1. Distribution of Sex, Injury Type, Explosive Type, Risk Activity, and Risk Knowledge by Age Group, Afghanistan—March 2001-June 2002 mines. Injuries from unexploded ord- No. (%) Injured, by Age Group, y* nance were more prevalent in children and adolescents younger than 16 years, Յ6 7-15 16-25 Ն26 Total (n = 86) (n = 664) (n = 453) (n = 433) (N = 1636) regardless of risk behavior. Because un- Sex exploded ordnance are more visible Male 69 (80.2) 593 (89.3) 437 (96.5) 400 (92.4) 1499 (91.6) than landmines, unexploded ordnance– Female 17 (19.8) 71 (10.7) 16 (3.5) 33 (7.6) 137 (8.4) related injuries are preventable and Status should become a priority in landmine Civilian 86 (100) 657 (99.0) 308 (68.0) 278 (64.2) 1329 (81.2) risk education efforts. Military 0 (0) 7 (1.0) 145 (32.0) 155 (35.8) 307 (18.8) Injury type The proportion of persons who re- Death 6 (7.0) 66 (9.9) 44 (9.7) 38 (8.8) 154 (9.4) ported knowing that the area was dan- Upper body 36 (41.9) 282 (42.5) 141 (31.1) 124 (28.6) 583 (35.6) gerous, that the area was marked for Lower body 27 (31.4) 94 (14.2) 131 (28.9) 139 (32.1) 391 (23.9) landmines and unexploded ordnance, or Upper and lower body 17 (19.8) 209 (31.5) 129 (28.5) 127 (29.3) 482 (29.5) that they had received mine risk educa- Unknown 0 (0) 13 (2.0) 8 (1.8) 5 (1.2) 26 (1.6) tion was low. However, using these re- Explosive type sults to evaluate mine risk education is Antipersonnel mine 14 (16.3) 125 (18.8) 208 (45.9) 223 (51.5) 570 (38.4) Antitank mine 3 (3.5) 16 (2.4) 13 (2.9) 11 (2.5) 43 (2.6) difficult because no estimates exist of the Cluster munition 5 (5.8) 62 (9.3) 27 (6.0) 23 (5.3) 117 (7.2) coverage of mine risk education in the Other unexploded ordnance† 56 (65.1) 397 (59.8) 165 (36.4) 143 (33.0) 761 (46.5) general population to permit calcula- Unknown/other 8 (9.3) 64 (9.6) 40 (8.8) 33 (7.6) 145 (8.9) tion of injury rates among persons who Activity at the time of injury did and did not receive such educa- Military activity 0 (0) 3 (0.5) 105 (23.2) 120 (27.7) 228 (13.9) tion. A rigorous evaluation of the effec- Traveling 6 (7.0) 53 (8.0) 65 (14.3) 62 (14.3) 186 (11.4) tiveness of landmine risk education in Collecting wood, food, or water 8 (9.3) 72 (10.8) 35 (7.7) 43 (9.9) 158 (9.7) reducing injuries due to landmines and Farming 3 (3.5) 35 (5.3) 36 (8.0) 52 (12.0) 126 (7.7) unexploded ordnance is needed in Af- Tending animals 11 (12.8) 171 (25.8) 57 (12.6) 23 (5.3) 262 (16.0) ghanistan and other countries. Playing/recreation 32 (37.2) 156 (23.5) 29 (6.4) 4 (0.9) 221 (13.5) The surveillance system detected an Tampering with explosive 4 (4.7) 58 (8.7) 21 (4.6) 15 (3.5) 98 (6.0) Other/unknown 22 (25.6) 116 (17.5) 105 (23.2) 114 (26.3) 357 (21.8) increase in injuries from cluster muni- Injured individual knew the area 4 (4.7) 38 (5.7) 59 (13.0) 72 (16.6) 173 (10.6) tions during October 2001, which coin- was dangerous cided with the start of bombings by coa- Injured individual received mine 4 (4.7) 77 (11.6) 34 (7.5) 68 (15.7) 183 (11.2) lition forces. This increase gradually risk education subsided after February 2002. These data *Percentages may not equal 100 due to rounding. †Includes grenades, bombs, and mortar shells. are consistent with reports about increased danger to civilians from unexploded cluster bomblets.6 The decrease being at highest risk. In addition, sur- volved in herding at a relatively young in such injuries may be attributed in part veillance data presented in this report age; this is reflected in our data, which to rapid clearance response measures concern only acute injuries and do not show that 13% of injured children aged undertaken by the Mine Action Center address long-term physical disability 6 years or younger and one quarter of for Afghanistan after receiving informa- and mental health concerns, which may injured children aged 7 to 15 years were tion from the US military about the loca- add substantially to both the eco- injured while involved in this activity. tion of coalition bomb strikes.5 nomic and public health burdens. These children may be at particularly The results of this study are subject to Furthermore, as our data show, be- high risk if herding prevents them from several important limitations. The ac- cause landmines and unexploded ord- attending school because many land- tual numbers of those injured and killed nance seriously hinder the simple ac- mine risk education programs tar- by landmines and unexploded ord- tivities of economic necessity, they may geted toward children are school- nance in Afghanistan are likely to be sub- substantially undermine postconflict based. Special programs focusing on stantially higher than that reported by recovery. The most important risk fac- children engaged in herding and not at surveillance data because the clinic- tors in adults (besides military activ- school may be necessary. based surveillance system is likely to miss ity) were activities of economic neces- Although prior reports and world those who die before reaching a clinic, sity: farming, tending animals, traveling, opinion often pay greater attention to those whose injuries are too minor to and collecting food, wood, and water. landmines, our data show that more seek medical care, and those living in However, this risk is not limited to than half of all injuries and deaths in areas with little or no access to health fa- adults. Afghan children begin to be in- Afghanistan were caused by unex- cilities that are involved in the surveil-

Downloaded From: https://jamanetwork.com/ on 09/25/2021 LANDMINES AND UNEXPLODED ORDNANCE

lance system. Because many fatal inju- Table 2. Distribution of Risk Activity by Explosive Type, Stratified by Age Group, ries are probably missed, a case-fatality Afghanistan, March 2001-June 2002 rate of 9.4% derived from surveillance No. (%) of Individuals Injured data is likely to be substantially under- Age and Activity estimated. Previous studies using popu- at Time of Injury Landmine* Unexploded Ordnance† Other/Unknown Total Age Ͻ16 y lation survey methods found a case- Military 2 (66.7) 1 (33.3) 0 (0) 3 12 fatality rate of 55% in Afghanistan, 41% Traveling 19 (32.2) 34 (57.6) 6 (10.2) 59 12 in Bosnia, 42% and 48% in Mozam- Collecting wood, food, or water 26 (32.5) 44 (55.0) 10 (12.5) 80 12,13 12 bique, and 31% in Cambodia. Farming 11 (29.0) 23 (60.5) 4 (10.5) 38 Time trends in injuries seen in the sur- Tending animals 55 (30.2) 102 (56.0) 25 (13.7) 182 veillance data should be interpreted with Playing/recreation 14 (7.4) 164 (87.2) 10 (5.3) 188 caution because of low sensitivity and Tampering with explosives 10 (16.1) 50 (80.6) 2 (3.2) 62 differential system coverage over time. Other/unknown 21 (15.2) 102 (73.9) 15 (10.9) 138 For example, the low number of inju- Total 158 (21.1) 520 (69.3) 72 (9.6) 750 ries recorded during September and Oc- Age Ն16 y tober 2001 may be due in part to dis- Military 183 (81.3) 29 (12.9) 13 (5.8) 225 ruption of surveillance activity before Traveling 84 (66.1) 34 (26.8) 9 (7.1) 127 and during the coalition strikes. Collecting wood, food, water 41 (52.6) 27 (34.6) 10 (12.8) 78 The disproportionate number of in- Farming 42 (47.7) 39 (44.3) 7 (8.0) 88 Tending animals 39 (48.8) 30 (37.5) 11 (13.8) 80 juries among men may be because Playing/recreation 8 (24.2) 23 (69.7) 2 (6.1) 33 women in Afghanistan are more re- Tampering with explosives 3 (8.3) 32 (88.9) 1 (2.8) 36 stricted than men in their mobility and Other/unknown 55 (25.1) 144 (65.8) 20 (9.1) 219 thus are less likely to engage in activi- Total 455 (51.4) 358 (40.4) 73 (8.2) 886 ties that put them at risk for landmine *Included antipersonnel and antitank mines. or unexploded ordnance injuries. In ad- †Included cluster munitions. dition, some of this observed sex dis- parity may result from decreased like- injuries, thus further increasing sensi- the risks of unexploded ordnance in- lihood of injured females to receive tivity. Using alternative methods for jury as much as landmine injury, par- medical care or to be interviewed and data collection, such as community- ticularly among children. recorded by the surveillance system. based surveillance or population- The absence of reliable data on the age based surveys would allow for record- Author Contributions: Study concept and design: distribution of the Afghan population did ing more representative data, including Bilukha, Brennan. Acquisition of data: Bilukha, Brennan, Woodruff. not allow for calculating age-specific in- fatal and minor injuries, and injuries Analysis and interpretation of data: Bilukha, Brennan. jury rates. However, by conservative es- among women and girls. Drafting of the manuscript: Bilukha, Brennan. timates it is unlikely that the group of 7- Critical revision of the manuscript for important in- Surveillance data suggest that land- tellectual content: Bilukha, Brennan, Woodruff. to 15-year-olds accounts for more than mine risk education should focus on Statistical expertise: Bilukha, Brennan, Woodruff. 25% of total population. Therefore, the landmine avoidance among adults and Administrative, technical, or material support: Bilukha, Brennan. finding that 40% of all injuries oc- avoidance of unexploded ordnance Study supervision: Bilukha, Brennan. curred in this age group suggests that this among children and adolescents. Sur- Funding/Support: This study was supported by funds from the Centers for Disease Control and Preven- group may indeed be at higher risk. veillance data from Iraq and other coun- tion. Expanding clinic-based surveil- tries where munition dumps have been Acknowledgment: We thank the International Com- mittee of the Red Cross for providing the data on in- lance to include nonparticipating health bombed may show a similar need to fo- dividuals injuried by landmines and unexploded ord- facilities would allow recording of more cus mine risk education programs on nance in Afghanistan.

REFERENCES 1. Centers for Disease Control and Prevention. Land- mine Monitor Report 2002. New York, NY: Human transfer of conventional weapons on civilian injuries: mine-related injuries, 1993-1996. MMWR Morb Mor- Rights Watch; 2002. retrospective analysis of prospective data from Red tal Wkly Rep. 1997;46:724-726. 6. Gall C. Farmers in Afghanistan face fields of bombs. Cross hospitals. BMJ. 1999;319:410-412. 2. Kakar F, Bassani F, Romer CJ, Gunn SW. The con- New York Times. October 9, 2002:A15. 10. Hanevik K, Kvale G. Landmine injuries in Eritrea. sequence of land mines on public health. Prehospital 7. Transitional Islamic Government of Afghanistan. Af- BMJ. 2000;321:1189. Disaster Med. 1996;11:2-10. ghanistan National Health Resources Assessment: Pre- 11. Jeffrey SJ. Antipersonnel mines: who are the vic- 3. Office of Humanitarian Demining Programs. Hid- liminary Results. Kabul: Transitional Islamic Govern- tims? J Accid Emerg Med. 1996;13:343-346. den Killers: The Global Landmine Crisis. Washing- ment of Afghanistan, Ministry of Public Health; 2002. 12. Andersson N, da Sousa CP, Paredes S. Social cost ton, DC: US Dept of State, Bureau of Political- 8. International Committee of the Red Cross Mine Ac- of land mines in four countries: Afghanistan, Bosnia, Cam- Military Affairs; 1998. tion Program. Semi-Annual Report ( January-June bodia, and Mozambique. BMJ. 1995;311:718-721. 4. Giannou C. Antipersonnel mines: facts, fictions, and 2002). Kabul, Afghanistan: International Committee 13. Ascherio A, Bielik R, Epstein A, et al. Deaths and priorities. BMJ. 1997;315:1453-1454. of the Red Cross; 2002. injuries caused by land mines in Mozambique. Lan- 5. International Campaign to Ban Landmines. Land- 9. Coupland RM, Samnegard HO. Effect of type and cet. 1995;346:721-724.

Downloaded From: https://jamanetwork.com/ on 09/25/2021