The Omagh bombing

On 15 August 1998, the largest single atrocity of the Northern Ireland conflict took place in Omagh, a market town with a population of 26,000, when a car bomb exploded in the town centre. Thirty-one people, including two unborn children (twins) were killed, 382 people were injured of which 135 were hospitalised. Twenty-six families were bereaved. Of those killed, 15 were aged 17 years or under. The bomb had a devastating effect on the community. A large number of those killed or injured were children and young people or adults with young families. Many children and young people sustained injuries resulting in the loss of limbs, loss of soft tissue, scarring and disfigurement. Many more were exposed to scenes of intense horror and suffering.

The first aim of this study was to assess the extent of psychiatric morbidity among children (aged 8 to 13 years) in a community following a car bomb explosion in the town centre on a busy Saturday afternoon. Children under eight were not included because of the different presentation of trauma reactions in these younger age groups [4]. Children and adolescents over the age of thirteen were included in another study to be reported at a later stage with more age appropriate measures. Secondly, we consider if type of exposure to a traumatic event increases PTSD symptoms in children to a greater extent than symptoms of general emotional distress. Thirdly, we investigate which individual and trauma characteristics identified within this study predict PTSD, depression and anxiety, and consider how our findings compare with the risk factors for PTSD in children and adolescents reported in Trickey and colleagues' recent meta-analysis [5] and other studies.

In relation to the first aim, most epidemiological studies have been of adults and older young people, such as the U.S. National Comorbidity Survey [6] that reported a 10% lifetime prevalence rate. In the U.K. National Mental Health Survey [7] a PTSD rate of 0.4% was found in children aged between 11-15 but scarcely registered below the age of 10 years. However the U.K. study reported a point prevalence estimate and the screening instrument used was not PTSD specific. Fletcher [8] in a meta-analysis of 34 studies reported that 36% of children who had experienced a range of traumas met criteria for PTSD. However, the rates of PTSD associated with traumatic events vary considerably from 0% to 100% [9]. In one review of natural disasters [10] 5-10% of children and adolescents met full criteria for PTSD and after road traffic accidents rates of 25 -30% have been recorded [11].

It has been established in many studies that increased exposure is associated with increased mental health problems including PTSD. In a review of 25 studies Foy and colleagues [12] found exposure to be one of three factors (severity of trauma exposure, trauma-related parental distress, and temporal proximity to trauma) that consistently mediated PTSD development in children. A relationship between level of exposure and PTSD has been found in studies of natural disasters [13–15] community violence [16, 17] and political conflict [18–20]. Higher PTSD rates have been reported in relation to specific characteristics of traumatic events, for example rates of 90% have been recorded following exposure to gruesome scenes [21]. In warfare studies of PTSD in children, incidence rates between 25% to 70% are reported depending on type of exposure and type of warfare [2, 22]. A number of studies have reported level of exposure and trauma severity as two main risk factors of PTSD [12, 23–25]. Trickey and colleagues [5] have identified trauma severity as the trauma characteristic most strongly associated with risk of PTSD in children and adolescents but suggest that trauma severity may be difficult to differentiate from trauma exposure. This poses the possibility of a range of psychological effects associated with a wider range of exposure categories including sub categories of direct exposure based on characteristics like proximity to the potentially traumatic event or being present at the time as opposed to just after an incident. Other established peri traumatic risk factors for PTSD such as physical injury [5], exposure to dead bodies [26] and perceived life threat [5] are theoretically more likely with more "direct" exposure such as being present at the time of a bombing compared with less direct exposure witnessing the immediate aftermath of a bomb. There is also evidence that other forms of indirect exposure such as exposure by media [27, 28] are linked to increased risk of PTSD. One concept that previous research does not appear to have systematically addressed is the psychological impact on children who are in the vicinity of an event such as a bomb but narrowly miss being at the precise location during or immediately after the event. We have defined this as a "Near Miss" category for analysis in this paper.

With respect to the third aim of this paper we consider how pre, peri and post trauma factors predict psychological reactions, particularly PTSD, in children following the Omagh bomb. In a recent comprehensive meta-analysis of risk factors for PTSD in children, Trickey and colleague's [5] reported risk factors for PTSD as follows: a small effect size for race and younger age; a small to medium-sized effect for female gender, low intelligence, low SES, pre and post-trauma life events, pre-trauma psychological problems in the individual and parent, pre-trauma low self-esteem, post-trauma parental psychological problems, bereavement, time post-trauma, trauma severity, and exposure to the event by media; and a large effect for low social support, peri trauma fear, perceived life threat, social withdrawal, comorbid psychological problem, poor family functioning, distraction, PTSD at time 1, and thought suppression.

In terms of pre-trauma factors, there have been contradictory findings from studies in relation to age [23, 29–31]. Trickey and colleagues [5] reported that younger age is largely unrelated to whether a young person develops PTSD but moderator analysis discovered that there was a statistically significant stronger relationship when the trauma was unintentional although the population effect size remained non-significant regardless of whether the trauma was intentional or non-intentional. Trickey and colleagues [5] also reported that younger age was a significant risk factor, with a small effect, if the index trauma was a group event rather than an individual one. There have also been conflicting findings regarding the relationship between gender and PTSD with some studies recording PTSD in girls at twice the rate as in boys [7]. Whist several studies have reported gender as a significant risk factor [12, 21, 24, 29, 32], Trickey and colleagues [5] reported female gender to be a consistent although statistically small risk factor and a stronger risk factor in older children and adolescents and also when the trauma is unintentional. Whilst girls seem more vulnerable to internalizing stress reactions, boys display more externalizing behaviour disturbance [24, 33]. Several studies have identified a number of pre-trauma risk factors including; prior traumas [20] prior psychiatric problems [25, 32, 34] and family cohesion [35]. Whilst type and severity of exposure are recognised as important predictors of PTSD in adults and children, studies have reported other specific peri-trauma factors including: a strong acute trauma response [23, 36, 37], witnessing dead people [26], being physically injured [10] and perceived life threat [24, 36, 37]. Post trauma factors associated with PTSD in children include: social support [25] and co-morbidity, especially depression and generalised anxiety [38–40].

Measures

Exposure to the bomb: Eight items covered various aspects of exposure to the bombing (see Table 8 in Appendix 1). On the basis of responses to these items, respondents were classified as belonging to one of five mutually exclusive exposure categories. “Exposed - in town at time" means was in Omagh town when the bomb exploded and witnessed injury or death of others or was directly harmed. “Exposed - in town after” means was in Omagh town shortly after the bomb exploded and witnessed injury or death of others or was directly harmed. “Loss” means did not witness injury or death of others, not injured but experienced loss or injury of someone close (family, relative or friend). “Near miss” means was in Omagh town when the bomb exploded but did not witness injury or death of others, was not directly harmed and did not experience loss. “No exposure” means was not in Omagh town when or after the bomb exploded, was not a witness and did not experience loss. In addition, children reported whether they had received any physical injuries (physically injured) or thought they were going to die (perceived life threat).

The Impact of Event Scale (IES) [42] is a widely used screening test for PTSD in children. In this study, the 8 item CRIES-8 (which lacks any arousal items) was used (α=0.82) as it was found to be as efficient as the CRIES-13 (which includes arousal items) in classifying children with and without PTSD [43]. It provides a continuous score for overall PSTD, and two sub-scales each consisting of four items: (1) intrusive thoughts, memories and images and (2) avoidance of thoughts and reminders. Items were grounded in the Omagh Bombing and referenced to experiences within the previous seven days.

The Birleson Depression Self-Rating Scale for Children (BDS) [44] is an 18-item scale assessing the level of depression in children (α=0.82). Items were scored on a three point scale (0,1,2). Responses include ‘most’, ‘sometimes’ and ‘never’. A score of 0 indicated a healthy response and a score of 2 indicated an unhealthy or depressed response.

The Spence Children’s Anxiety Scale (SCAS) [45] consists of 38 items on specific anxiety symptoms with a further six filler items (α=0.94). Responses include ‘never’, ‘sometimes’ , ‘often’ and ‘always’ and are recorded on a four-point scale (0,1,2,3). The scale provides a global anxiety rating together with scores on six individual subscales covering specific anxiety symptoms, namely separation anxiety, social phobia, obsessive-compulsive disorder, panic/agoraphobia, generalised anxiety, and, fears of physical injury.

Socio-demographics: Each respondent provided details of their age and gender, as well as information on family structure (living with both parents/living with single parent/reconstituted family/in state or foster care) and parental employment (both parents employed/mother employed and father not employed/father employed and mother not employed/both parents not employed) (Table 1). Post event support was measured by asking if help was received because of difficulties experienced following the bomb and a checklist of sources of help was provided to identify the provider(s).

Statistical analysis

A series of OLS regression models were estimated to examine the predictors of PTSD, anxiety and depression. A three step hierarchical regression was conducted with the predictor variable included in blocks corresponding to pre-, peri- and post-trauma variables. These models were restricted to those individuals who were in town on the day of the bombing and/or witnessed traumatic events. As the sample was clustered at the school level, school dummy variables were included in the model to account for the lack of independence due to school clustering. This ensures that the regression standard errors are adjusted for the lack of independence at the school level. While these dummy variables were included within the model they were not reported within the presented regression tables. None of the school level dummies were significant within the various models.

Psychiatric morbidity

Type of exposure: associations with PTSD and other psychiatric disorders

No age or gender variations were noted across the levels of exposure (Table 4). The mean scores on the IES, BDS and the SCAS were 15.65, 8.67 and 27.42 respectively (Table 1). The PTSD, depression and anxiety scores varied significantly across types of exposure, with increased exposure associated with higher scores on the IES, BDS and SCAS (Table 4). There were significant differences between the level of exposure and PTSD symptoms (F(4,1856)=37.698, p<0.01), depression (F(4,1867)=8.138, p<0.01) and anxiety (F(4,1778)=18.179, p<0.01). Figure 1 shows the IES, SCAS and BDS standardised symptom scores for each type of exposure. An increase in level of exposure is associated with increased levels of PTSD. However, those in the near miss group exhibited higher levels of anxiety and depression than the loss group. Direct exposure (those present at the time of the explosion and those present after the explosion) was associated with larger increases for PTSD symptoms than for general psychiatric distress. Paired comparisons of these differences showed that standardised IES scores of the two groups directly exposed differed significantly compared to the loss group (p<0.01), no exposure (p<0.01) and the near miss group (p<0.05). The differences between the two groups directly exposed to the bomb scenes, (those present at the time of explosion and those present after the explosion) were not significant on the IES (p=0.255), SCAS (p=0.663) and depression measures (p=0.604). The anxiety scores (SCAS) of those in the two exposure groups were significantly different to those in the no exposure (p<0.01), loss (p<0.01) groups but not the near miss (p=0.334) group. On the depression measure (BDS) those directly exposed differed significantly compared to the loss group (p<0.01) and the no exposure group (p<0.01), but not the near miss group (p=0.494).

	Type of exposure					p
Characteristic	No exposure	Near miss	Loss	Exposed - in town after	Exposed - in town at time
Age (Mean)	11.5	11.3	11.4	11.3	11.2	0.233
IES (Mean)	13.17	16.05	17.07	20.29	21.73	0.00**
BDS (Mean)	8.08	9.35	8.92	10.02	10.38	0.00**
SCAS (Mean)	24.23	30.37	29.19	33.85	34.89	0.00**
Female (%)	50.7	55.0	51.3	53.8	52.5	0.959
Previous psychological treatment (yes)	1.3	0.0	3.6	6.9	8.2	0.00**
Post-event support (yes)	0.7	0.0	1.9	7.0	15.0	0.00**
Perceived life threat (yes)	-	0.0	-	6.0	21.1	0.003**
Physically injured (yes)	-		-	10.9	9.9
Witnessed serious injury (yes)	-		-	92.1	98.0
Witness people dying (yes)	-		-	61.1	70.3
Witnessed people dead (yes)	-		-	39.8	57.4

Note: “Exposed - in town at time" means in Omagh town when the bomb exploded and witnessed injury or death of others or was directly harmed. “Exposed - in town after” means in Omagh town shortly after the bomb exploded and witnessed injury or death of others or was directly harmed. “Near miss” means in Omagh town when the bomb exploded but did not witness injury or death of others and was not directly harmed. “Loss” means experienced loss or injury of someone close (family, relative or friend) but no direct harm. “No exposure” means not in Omagh town at the time or shortly after the bomb exploded, not a witness and did not experience loss. “IES”: Impact of Events Scale; “BDS”: Birleson Depression Scale: “SCAS”: Spence Children’s Anxiety Scale. P: *significant at the 0.05 level; **significant at the 0.01 level.

Predictors of PTSD and other psychiatric disorders

Of those directly exposed to the bomb approximately one in ten received post-event psychological/psychiatric interventions. Post-event support significantly predicted probable anxiety (p<0.01; Table 6) but not PTSD (Table 5) or depression (Table 7). Those who received support had significantly higher levels of depression (t(216)=3.007, p<0.01) and anxiety (t(201)=3.656, p<0.01). However, no significant differences in PTSD scores were observed (p=0.057).

Limitations

Our data was gathered 15 months after the bomb so it is likely that screening in the immediate aftermath of the bomb would have identified higher levels of PTSD symptomotology. Our questionnaire did not capture any traumas or significant life events that children may have been experienced in the intervening period that may have compounded an initial traumatic reaction to the bombing. Self-report questionnaires were used in the screening and we recognise these are only an indicator of probable psychiatric disorders and do not provide a complete accurate diagnosis. We were unable to collect multi-informant data from parents or teachers which would have provided confirmatory data to identify morbidity amongst the sample. While the overall sample size was large, the number of children who were directly exposed to the bombing was relatively small. This will have reduced the statistical power of the regression models. Finally, the study assessed psychological symptoms but did not measure the impact of symptoms on daily functioning.