Stress Reactions of Children and Adolescents in War and Siege Conditions

Eighty percent of the victims of war are reported to be children and women (1). The psychological impact of war on them has been studied in World War II (2) and, more recently, in wars in Israel, Lebanon, and Kuwait (3). Saigh (4,5) reported that 32% of children aged 9–13 years developed posttraumatic stress disorder (PTSD) during and after the Beirut conflict. Chimienti et al. (6) estimated that the children exposed to the armed conflict in Lebanon had 1.7 times more symptoms of PTSD than the general population. A high incidence of comorbid depression with PTSD has also been reported (7–9).

The effect of long-standing siege on children is less researched because of the rare occurrence of siege conditions these days and the difficulties those conditions present for research. This study is unique in that the participants lived in prolonged siege conditions and were deprived of basic necessities of life for over 1 year. They were exposed to random sniper fire from the hills surrounding the city of Sarajevo, and many lost family members and close friends.

This study explored the relationship of gender, loss of family members, and perceived deprivation to the development of posttraumatic stress symptoms.

METHOD

Because of the unusual circumstances that the siege of Sarajevo presented at the time of the study, the Institutional Review Board of the University of Missouri School of Medicine accepted verbal consent from parents or guardians of children who participated in the study. The principal of the children’s school explained the purpose of the study to the parents and the children and requested permission for participation, which none declined.

Data were collected in February and March 1994 from all students (ages 7–15 years) in one school district. We could not determine how our sample differed from children in other school districts in Sarajevo. Data on socioeconomic status are not included because of the widespread deprivation resulting from the war.

All scales were translated into the Bosnian language by a bilingual doctoral-level Bosnian psychologist. The Bosnian versions were independently retranslated into English and compared with the originals for accuracy. The scales included the Children’s Posttraumatic Stress Reaction Index (10), which measures PTSD symptoms and includes items such as intrusive imagery, poor concentration, and bad dreams; the Impact of Event Scale (11), which measures the subjective emotional response after stressful life events; the Children’s Depression Inventory (12), which measures current level of depression; and a general information questionnaire, which included demographic information pertaining to loss of family members, displacement, and perceived needs.

Since neither the Children’s Posttraumatic Stress Reaction Index nor the Impact of Event Scale alone fully conforms with the DSM-IV criteria for PTSD, the information from the two scales was combined to form a “composite DSM-IV PTSD scale,” which elicited information regarding experience of an extreme stressor and symptoms of avoidance, hypervigilance, and reexperiencing. Since every facet of life was in disarray because of the war, level of functioning could not be assessed.

Chi-square analyses determined the independence of categorical variables. Wilcoxon signed-ranks tests compared middles of distributions of continuous variables.

RESULTS

Of the 791 children studied, data on age were lacking for 70 and data on gender for 11. The mean age was 11.0 years (SD=2.3). The age distribution was as follows: 253 (32%) of the children were 7–9 years old, 236 (30%) were 10–12 years old, and 232 (30%) were 13–15 years old. Forty-nine percent of the children were male, and 51% were female.

For the girls, the mean score on the avoidance items of the Impact of Event Scale was 23.6 (SD=6.9), the mean score on the reexperiencing items was 15.1 (SD=5.2), and the mean total score was 38.6 (SD=11.1). For the boys, these values were 21.6 (SD=6.8), 13.8 (SD=5.0), and 35.3 (SD=10.8), respectively. The differences between the girls and boys in all three values were significant (z=–4.37, p=0.0001; z=–3.54, p=0.0004; and z=–4.51, p=0.0001, respectively).

Of 719 subjects, 613 (85%) had experienced direct or indirect sniper fire; data were missing for 72 children who did not respond to the question about this. There were no significant differences in the scores on the various scales of those who reported sniper fire and those who did not.

A total of 521 (66%) of the children had lost a member of the immediate or extended family; children who had experienced such a loss had more symptoms of PTSD. Twenty-six percent of our sample experienced food deprivation, 48% experienced clothing deprivation, 29% reported shortage of water, and 10% reported lack of shelter. The children who experienced lack of water and shelter were significantly more likely to manifest avoidance and reexperiencing symptoms than those who did not (table 1).

DISCUSSION

Our findings are consistent with the published literature (13) in that more girls than boys reported avoidance behaviors and reexperiencing symptoms. However, the differences are small, even though significant, and should be interpreted with caution.

Eighty-five percent of our sample had experienced sniper fire, but there was no significant relationship between experiencing sniper fire and the development of PTSD symptoms.

The loss of a family member and deprivation of food, water, and shelter had a severe adverse impact on the children. They identified the needs for food and clothing more frequently than the needs for water and shelter. In general, deprivation was associated with significantly increased symptoms of avoidance and hypervigilance. These results imply that proximity to war atrocities and personal losses are highly correlated with the development of symptoms of PTSD.

Received Feb. 17, 1998; revision received June 22, 1998; accepted July 16, 1998. From the Department of Psychiatry and Neurology, the Department of Education, and the Department of Engineering, University of Missouri-Columbia. Address reprint requests to Dr. Husain, Department of Psychiatry and Neurology, School of Medicine, University of Missouri, Columbia, MO 65212. Supported in part by grant 95-RC-071-ER from the Research Council of the University of Missouri-Columbia, grant RR-07053 from NIH, and grant NSF CMS 9411866 from the National Science Foundation and the Council for Humanitarian Agencies.The authors acknowledge the contribution of the late Dennis P. Cantwell for his input and fruitful discussions related to this study.

TABLE 1

Enlarge table