The development of posttraumatic stress disorder (PTSD), as defined by current psychiatric classification systems, is contingent on exposure to a traumatic event. Yet, only approximately 10% to 20% of people who experience trauma subsequently develop the disorder (1). Consequently, the currently held view is that the event itself is only one of several determinants responsible for the transition from a normal response to PTSD (1,2). Accordingly, considerable research has been devoted to the identification of predictors and risk factors that may enhance the likelihood of developing PTSD after exposure to a traumatic event.
Despite the large body of research on predictors and risk factors for PTSD, a comprehensive understanding of the development of the disorder remains elusive. In part this is the result of the methodological constraints inherent in the study of trauma such as the difficulty in identifying an appropriate sample of trauma survivors before traumatic exposure. As a result, most studies in the field have used retrospective or cross-sectional designs with convenience samples drawn from chronic populations seeking psychiatric treatment. Conclusions from such studies are therefore limited (1–3).
Moreover, it has become increasingly clear that the consequences of exposure to trauma are affected not only by aspects of the event itself, but also by aspects of the survivor's reaction to the event (4,5). Accordingly, research has been directed toward the investigation of factors that may have an impact on the trauma response such as personality traits, the perceived threat posed by the event, and the individual's coping style by which adverse aspects of the environment are modified and the internal threat induced by stress or trauma can be minimized (4–6).
A recent study of Palestinian former political prisoners (4) found that the acuteness of trauma, the subjective evaluation of the traumatic experience as harmful and involving loss, and the use of both emotion- and problem-focused coping efforts were all associated with the severity of PTSD symptoms. Problem-focused coping, defined as efforts to recognize, modify, or eliminate the impact of a stressor (5,6), was found to be associated with lower levels of PTSD symptoms at the outset. Conversely, emotion-focused coping, defined as efforts to regulate emotional states that are associated with exposure to stress (5,6), was associated with lower levels of PTSD symptoms in the long run. The retrospective design of this study, however, does not allow for the classification of coping style as a pretraumatic predictor for PTSD.
Indeed, to successfully determine pre- and posttraumatic risk factors and predictors for PTSD, research necessitates the use of prospective designs that include both pre- and postexposure assessments. One such rare study (7) examined the personality profile of undergraduate students before a traumatic exposure, and once more, decades later, after combat exposure. The findings illustrated that high levels of neuroticism and introversion predicted the development of PTSD. Yet, the prolonged period between the initial and final evaluations in this study prevented the necessary control of potential confounding variables that might also have had an effect on the outcome above and beyond the impact of trauma.
The present study examined the role of pretraumatic personality factors and coping style in the prediction of PTSD following a suicide bomber's attack on a bus. The unique and tragic circumstances of this event in which 17 passengers were killed and over 30 others were physically injured enabled the utilization of a prospective design and the exploration of the impact of the proximity to the attack and its perceived threat on the response of survivors.
Sample and Procedure
The initial sample consisted of 185 first-year undergraduate students from the Faculty of Social Welfare and Health Studies at the University of Haifa in Israel. The initial evaluation was conducted by the corresponding author during an introductory psychology class for the purpose of identifying predictors of academic achievement. The evaluation, consisting of the assessment of personality traits and trait coping style, took place in the spring of 2003, coincidentally 2 weeks before a suicide bomb explosion on a bus that was en route to the university. After approval by the university's human research review committee and receipt of informed consent from the participating students, the focus of the study was modified from the measurement of academic achievement to a prospective study of the posttraumatic stress response.
Multiple assessments were conducted, as detailed in the research plan indicated in Table 1.
One week after the attack, the participants were questioned about the extent of their actual exposure to the bus explosion (proximity to the attack) and the perceived threat posed by the attack. One month after the attack, the participants were reevaluated to assess potential changes in personality traits and coping style. Six months after the attack, a formal diagnostic assessment of PTSD was conducted.
Of the 185 students present in the classroom during the initial evaluation, two (1%) refused to participate in the research and three (1.6%) were excluded from participation as a result of language limitations. The final cohort of 180 subjects had a mean age of 23.4 (standard deviation [SD] = 4.7) years; 12.6 (SD = 1.5) years of schooling; and were predominantly female (N = 104 [58%]), single (N = 137 [76%]), and Israeli-born (N = 122 [68%]). Eighty (44%) of the participants were enrolled in the Department of Nursing, 60 (33%) in the Department of Human Services, and 40 (22%) in the School of Social Work. Nine (5%) had a history of psychiatric disorders: three with a diagnosis of PTSD, five with a diagnosis of social phobia, and one with a diagnosis of mood disorder.*
Proximity to the attack was evaluated by the following question: “Please select and mark the one item listed below that best describes your involvement in the recent terrorist explosion in Haifa on bus number 37: 1) I was on the bus; 2) I witnessed the explosion in person (that is, not on television or in the newspaper); 3) significant others (family members or close friends) were killed in the explosion; 4) significant others (family members or close friends) were injured in the explosion; 5) significant others witnessed the explosion and shared its details with me; and 6) I was not involved in any way in the bus explosion.” Six students (3.3%) reported being inside the bus, 25 (13.9%) witnessed the explosion in person, one (0.6%) had a significant other who was killed in the explosion, 13 (7.2%) had significant others who were injured in the explosion, and 36 (20.0%) had significant others who witnessed the attack and shared these details with them. Ninety-nine students (55.0%) reported no involvement in the attack.
As a result of the small number of respondents who were on the bus or had significant others who were killed, the proximity to the attack was transformed into a variable with three categories: direct exposure, including a positive reply to 1) and 2), reported by 31 (38%) respondents; indirect exposure, consisting of a positive reply to 3) and 5), reported by 50 (62%) respondents; and no exposure, consisting of a positive reply to 6), reported by 99 (55%) respondents.
Personality traits were assessed by the Tridimensional Personality Questionnaire (TPQ) (8), a 100-item true/false self-report inventory designed to measure the three primary personality dimensions thought to be involved in the mediation of particular types of stimuli: 1) novelty-seeking (NS)—a tendency to respond with intense excitement to novel stimuli or to potential relief from punishment (34 questions); 2) harm avoidance (HA)—a tendency to respond fearfully, tensely, and with inhibition to previously established signals of adverse stimuli and to learn passively to avoid punishment, novelty, and nonrewarding situations (33 questions); and 3) reward dependence (RD)—a tendency to respond with oversentimentality and with particular sensitivity to signals of reward, especially social reward, and to maintain behavior previously associated with reward or with relief from punishment (33 questions). The questionnaire was found to be valid and useful in studies of various psychiatric disorders (9), including PTSD (12). The total score of each of the three dimensions is calculated as the sum of the participants' ratings. The Hebrew version of the TPQ, used in the present study, was found to have strong reliability and validity in previous studies (8,9) as well as in the present sample (Cronbach alphas ranged from 0.85 for reward dependence in the preattack evaluation to 0.90 for novelty-seeking in the postattack evaluation). Reliability was further demonstrated by the finding that the intercorrelation between the three subscales was insignificant both in the present study and in the initial TPQ validation study (8).
Trait and state coping styles were examined by the Multidimensional Coping Inventory (COPE) (10), an instrument consisting of 60 statements describing 13 distinct ways of responding to stressful situations with self-evaluations ranging from zero (not at all) to 3 (very much so). Five of the statements (relating to active coping, planning, suppression of competing activities, restraint coping, and seeking instrumental social support) describe problem-focused coping; five other statements (relating to seeking emotional social support, positive reinterpretation, acceptance, denial, and turning to religion) describe emotion-focused coping. Three statements (focusing on venting emotions, behavioral disengagement, and mental disengagement) describe avoidance, which is a less useful coping style. In the initial assessment (T0), participants were asked to rate the correspondence of each item with their own typical reaction to general everyday stress (trait coping style). In the second evaluation (T1), participants were requested to relate their responses on the 60 items to the way they were dealing with the suicide bombing. The categorization into three coping styles, described previously, was accomplished by summing up the participants' responses to each statement. Internal consistency for the three coping styles, computed from baseline data, ranged from Cronbach alpha values of 0.90 (trait avoidance coping style) to 0.93 (state problem-focused coping style).
Perceived threat—One week after the attack, all subjects in the study were asked to rate the degree to which they experienced the bus explosion as a personal threat to their physical or emotional well-being on a scale ranging from zero (did not experience the explosion as a threat at all) to 10 (experienced the explosion as a major threat).
A formal diagnostic assessment of PTSD was conducted 6 months after the attack with Hebrew version of the Structured Clinical Interview for Axis I DSM-IV Disorders (SCID) (11). This semistructured interview schedule, which targets past and current major axis I diagnoses, is designed to be administered by a clinician or a trained mental health professional familiar with the DSM classification and diagnostic criteria.
The main research issues underlying the analyses were the role of preattack personality traits, preattack coping style, proximity to the attack, and perceived threat in predicting the development of PTSD 6 months after the bus explosion. First, the sample was divided into two groups according to the presence or absence of PTSD. The two groups were then compared in terms of demographic profiles by χ2 and t test analyses to identify potential confounding background variables that play a role in the development of PTSD (Table 2).
As shown in Table 2, 6 months after the traumatic exposure, 31 participants (17%) met the full criteria for a diagnosis of PTSD, whereas 149 (83%) did not. The two groups appeared to be similar in age, gender distribution, marital status, and country of origin. The mean years of education reported by the sample was 12.6 (SD = 1.5).†
Statistically significant differences were found only in level of education (t = 1.9, df = 178, p < .05).
Associations between each of the four independent variables—proximity to the attack, personality traits (TPQ scores), coping style (COPE scores), and perceived threat with PTSD—were examined separately at the different points of assessment as shown in Table 3.
Proximity to the Attack
Initially, participants in the three categories of proximity to the attack (i.e., direct exposure, indirect exposure, and no exposure) were compared in terms of the perceived threat posed by the attack. The results of an analysis of variance (ANOVA) followed by Duncan post hoc suggested that proximity was not associated with degree of perceived threat (F = 3.1, df = 178, p > .05), thus supporting the inclusion of the 99 participants who reported no exposure to the attack in subsequent analyses. In addition, as indicated by a χ2 analysis (χ2 = 29.3, df = 178, p < .001), proximity to the attack was significantly associated with PTSD so that the crude relative risk of developing PTSD 6 months after the attack for participants directly exposed to the bus explosion was almost five times higher than for those who were exposed indirectly and for those who reported no exposure (odds ratio [OR] = 4.9; confidence interval [CI] = 1.0–10.9; p < .001).
Repeated-measure multivariate analysis of variance revealed a significant effect for group (with/without PTSD) in NS (not significant; F = 17.3, df = 178, p < .01) and HA (F = 17.1, df = 178, p < .01) personality dimensions. No statistically significant effect was found for time of assessment or for the interaction between group and time. ANOVA followed by Duncan post hoc showed that subjects with PTSD appeared to score lower on the NS personality dimension and higher on the HA dimension compared with those without PTSD both before and after the attack.
Because the assessment of coping style 2 weeks before the attack was directed at assessing trait coping style, whereas the assessment of coping style 1 month after the attack was directed at assessing state coping style, the initial analyses focused on a comparison between the subjects' pretrauma trait coping style and posttrauma state coping style. Using a series of t test analyses, the two groups were compared with regard to each cross-sectional assessment, suggesting that participants with PTSD scored lower on both the trait emotion-focused coping style and the state problem-focused coping style (t = 2.2, df = 178, p < .001; t = 2.1, df = 178, p < .001, respectively) and higher on the trait and state avoidance coping styles (t = 2.1, df = 178, p < .001; t = 1.9, df = 178, p < .001, respectively). No statistically significant differences were observed between the two groups in terms of trait problem-focused coping style and state emotion-focused coping style. More specifically, participants with PTSD scored lower on the emotion-focused coping style before the terrorist attack and lower on the problem-focused coping style after its occurrence. In addition, participants with PTSD scored higher on the avoidance coping style both before and after the attack as compared with those without PTSD.
The level of perceived threat posed by the bus explosion was also significantly associated with PTSD: participants with PTSD reported higher levels of perceived threat as compared with those without PTSD (t = 2.6, df = 178, p < .001).
Lastly, a hierarchical model was examined using a series of logistic regression analyses with the aim of predicting PTSD (no/yes) 6 months after the attack (Table 4). Variables were entered in the order of the strength of their relationship to PTSD found in the preceding analyses. Step 1 contained the three categories of proximity to the attack: direct exposure, indirect exposure, and no exposure. Step 2 added the three preattack personality dimensions: NS, HA, and reward dependence. Step 3 added the three types of trait and state coping styles: problem-focused, emotion-focused, and avoidance. Step 4 added the perceived threat posed by the attack. To account for differences in level of education, this variable was entered into the regression analysis as a covariate confounder. The criterion for entry and retention in the regression equations was a significance level of p < .01. As shown in Table 4, the final model revealed six predictors that together accounted for 37% of the variance in the prediction of PTSD at 6 months (R2for step 4 = 0.37, p < .001). Specifically, once all six variables in this sample were entered in the model, an increased risk for PTSD was associated with direct exposure (OR = 4.10; CI = 1.30–5.90; p < .001), indirect exposure (OR = 1.40; CI = 1.00–1.90; p < .001), and preattack HA personality dimension (OR = 1.32; CI = 1.12–1.90; p < .01). State avoidance coping style (OR = 1.13; CI = 1.09–1.20; p < .01) and perceived threat (OR = 1.12; CI = 0.70–1.90; p < .05) were found to have a low association with increased risk for PTSD.
The tragic circumstances that led to the present study were caused by a suicide bomber's explosion on a bus traveling on a busy route to the University of Haifa, Israel, at midday, a time elementary schoolchildren are on their way back from school. This incident was one of several in a single year (2003) that shocked the residents of Haifa, a city whose population is 10% Arab, with a long record of peaceful coexistence.
The undergraduate students who participated in the research had been coincidentally assessed 2 weeks before the bus explosion for the prediction of academic achievement. Once the explosion occurred, the corresponding author repeated assessments at 1 week, 1 month, and 6 months afterward, thus enabling a prospective examination of the relationship of preattack personality, preattack coping style, and factors related to the traumatic event and to the individual's immediate response—often referred to as peritraumatic factors—to the development of PTSD (2).
The primary findings of the study demonstrate that the specific personality dimension of HA and the perception of personal threat posed by the attack are indeed meaningful risk factors for the development of PTSD after exposure to a traumatic event. The independent role of these two factors in predicting PTSD is reinforced by the fact that no interaction effect was observed between them.
Participants who demonstrated high levels of the HA personality dimension before the attack were found to be at increased risk for the development of PTSD at 6 months. This finding largely concurs with previous studies demonstrating an association between the HA dimension and various psychiatric states (9), including increased risk for PTSD symptoms (12). Richman and Frueh (12), who also used the TPQ in the context of PTSD, found an association between all three personality dimensions and the severity of posttraumatic stress symptoms (PTSS) among Vietnam veterans. However, because their study used a retrospective design, the comparison is limited.
Considering that HA appears to be closely related to high neuroticism and low extroversion and that both personality traits have been associated with PTSS and PTSD (13), the present study provides preliminary support for the possibility that the HA dimension may be considered a personality indicator for the development of PTSD. Individuals with high HA tendencies were described by Cloninger (8) as cautious, worriers, inhibited, tense, and apprehensive as well as intensely responsive to previously established signals of aversive stimuli. Identifying individuals who can be characterized as having these tendencies may enable early intervention and consequent prevention of PTSD in the immediate aftermath of exposure to trauma. However, because this study is preliminary and exploratory, the findings require replication before such conclusions can be drawn. Moreover, the ongoing exposure to terrorist attacks in Israel during the months before and after the study is another factor that warrants consideration, because it may be responsible for the heightened HA tendencies in the preattack assessment. Indeed, it is conceivable that participants were already at higher risk for PTSD before exposure to the bus explosion even without actual involvement in previous traumatic events. However, the premise is debatable in view of the finding that the participants' HA scores did not change significantly after the terrorist attack.
Supporting most research in the field (13,14), including findings reported by Carver and associates, whose stress theory formed the basis for the coping-style questionnaire used in the current study (10), our findings demonstrate that the state avoidance coping style is a significant predictor of the development of PTSD. Specifically, high levels of the avoidance coping style increase the risk for the development of PTSD at 6 months. Given that this coping style is characterized by emotional venting and behavioral and mental disengagement, the findings reinforce the widespread approach to PTSD as a manifestation of the failure to cope with traumatic thoughts and memories through an inefficient coping style. Previous research has shown that patients with PTSD tend to invest enormous efforts in suppressing the impact of the traumatic event but that these efforts are generally futile, because this type of coping is highly associated with intrusive traumatic thoughts and avoidance tendencies (13).
Another important finding of the study is that direct exposure to the attack significantly increased the risk for the development of PTSD at 6 months. Supporting previous literature (2), this finding indicates that type of exposure to traumatic events can be considered a significant risk factor for PTSD, namely the more direct and personal the exposure, the greater the risk for PTSD. However, although direct exposure to the attack was found to be the strongest predictor for PTSD, the findings of the study suggest that indirect exposure to the attack was associated with an increased risk for the development of PTSD as well. This result supports the observations that underlie the concept of secondary/vicarious traumatization (15) and the notion that the perception of threat (i.e., counterfactual thinking such as “I could have been on that bus”; “The university is no longer a safe place”) may reinforce the traumatic experience as suggested by the introduction of the subjective appraisal of the event in criterion A.2 of the DSM-IV definition of PTSD (16).
Several methodological limitations in the study should be noted. First, all the participants were university students and thus had a higher level of education than the general population. Moreover, the participants were predominantly female students at the departments of nursing and social work. Although subsequent analyses indicated that gender was not associated with the risk of developing PTSD, the choice of profession may suggest a bias specific to the current sample. However, the first year of schooling does not include practical training of the kind that may have provided these students (compared with students enrolled in human resources studies) with tools that enhance their ability to cope with trauma. We are therefore of the opinion that although the generalization of the results should not be affected by the distribution of gender and study areas, it is possible that individuals with certain personality features that do characterize trainees in the “helping” professions may also influence coping with trauma or resilience in ways not examined in the current study. Lastly, the participants' history of exposure to traumatic events, a well-known risk factor for PTSD, was not assessed in the study and its relative contribution to the prediction of PTSD is thus undetermined.
Notwithstanding the limitations of the study, its prospective nature provides a rare opportunity to discern the mechanisms that heighten the vulnerability of certain individuals, as described previously, to develop PTSD after exposure to a shocking event such as a terrorist attack. The findings of the study further demonstrate the important and informative role, in the subsequent development of PTSD, of characteristics and attributes that relate to the traumatic event itself and to the subjective appraisal of perceived threat.
1. Breslau N, Kessler R, Chilcoat H, Schultz L, Davis G. Trauma and posttraumatic stress disorder in the community: the 1996 Detroit Area survey of trauma. Arch Gen Psychiatry 1998;55:626–32.
2. McFarlane AC. Posttraumatic stress disorder: a model of the longitudinal course and the role of risk factors. J Clin Psychiatry 2000;61(suppl 5): 15–23.
3. Ballenger JC, Davidson JR, Jonathan RT. Consensus statement on posttraumatic stress disorder from international consensus group on depression and anxiety. J Clin Psychiatry 2000;61:60–6.
4. Kanninen K, Punamaki R, Qouta S. The relation of appraisal, coping efforts, and acuteness of trauma to PTS symptoms among former political prisoners. J Trauma Stress 2002;15:245–53.
5. Lazarus S. Emotion and Adaptation. New York: Oxford University Press, 1991.
6. Folkman S, Lazarus R. Coping and emotion. In: Monat A, Lazarus R, eds. Stress and Coping: An Anthology. New York: Columbia University Press, 1991:207–27.
7. Schnurr PP, Friedman MJ, Rosenberg SD. Preliminary MMPI scores as predictors of combat-related PTSD symptoms. Am J Psychiatry 1993;150:479–83.
8. Cloninger CR. A systemic method for clinical description and classification of personality variants. Arch Gen Psychiatry 1987;44:573–87.
9. Gil S. The role of personality traits in the understanding of suicide attempt behavior among psychiatric patients. Archives of Suicide Research 2003;7:1–8.
10. Carver S, Scheier M, Weintraub K. Assessing coping strategies: a theoretically based approach. J Pers Soc Psychol 1989;56:267–83.
11. First MB, Spitzer RL, Gibbon M, Williams J. Structural Clinical Interview for Axis I DSM-IV Disorders (SCID). New York: Biometrics Research Department, 1994.
12. Richman H, Frueh BC. Personality and PTSD II: personality assessment of PTSD-diagnosed Vietnam veterans using the Cloninger Tridimensional Personality Questionnaire (TPQ). Depress Anxiety 1997;6:70–7.
13. Fauerbach JA, Lawrence JW, Schmidt CW, Munster AM, Costa PJ. Personality predictors of injury-related posttraumatic stress disorder. J Nerv Ment Dis 2000;188:510–77.
14. Jarle E. The course of PTSD symptoms following military training accidents and brief psychological intervention. Personality and Individual Differences 2003;35:771–83.
15. Figley C. Helping Traumatized Families. San Francisco: Jossey-Bass, 1989.
16. Diagnostic and Statistical Manual of Mental Disorders, Fourth Revised Ed (DSM-IV-TR). Washington, DC: American Psychiatric Association, 2001.
* The exclusion of participants with a history of psychiatric disorder from subsequent analyses had no impact on the results, thus supporting the continuous use of the entire sample (N = 180). Cited Here...
† This range is most likely the result of enrollment in a 1-year preparatory program offered in advance of the undergraduate program. Cited Here...