Individual Predictors of Posttraumatic Distress: A ... - SAGE Journals

Individual Predictors of Posttraumatic Distress: A Structural Equation Model Cheryl Regehr, PhD1, David Hemsworth, MBA2, John Hill, BA, BSc3

Objective: Recent research has called into question the “dose-effect” model of understanding response to trauma and has turned attention to the contribution of personality and environmental factors. This research seeks to model the interrelation of relational capacity (a component of personality), perceptions of social support, and posttraumatic distress. Method: A group of firefighters (n = 164) completed questionnaires that addressed exposure to traumatic events, social support, current level of distress, and relational capacity. Structural equation modelling was used to develop a framework for understanding traumatic reactions. Results: The overall fit of the hypothesized model was excellent. Relational capacity had a significant negative effect on support, indicating that perceived social support decreased as disturbances in relational capacity increased. Perceived social support had a significant negative effect on level of distress. Conclusion: While some emotional response to disturbing events may be normal, the severity of symptoms covaries with the ability of the individual to develop and sustain supportive relationships to buffer the impact of events. (Can J Psychiatry 2001;46:156 – 161) Key Words: critical incidents, trauma, relational capacity, social support, attachment

B

y acknowledging the contribution of environmental stressors to mental health problems, the emergence of the diagnostic category posttraumatic stress disorder (PTSD) heralded an important shift in psychiatry. Implicit in the criteria for diagnosis was the assumption that PTSD could occur in any individual who encountered a catastrophic event, regardless of any pretraumatic considerations (1). This assumption has been supported by a body of research literature which concluded that the development of posttraumatic distress is correlated with the dosage of traumatic exposure (2–4). Other authors, however, argue that the dose-effect model inadequately explains posttraumatic distress. These critics point to methodological limitations of dose-effect studies, small amounts of variance explained by the severity of the traumatic event, and low incidence rates of PTSD following shared events (5–7). While trauma severity undoubtedly contributes to the distress experienced by individuals, it is becoming increasingly clear that trauma and distress do not have a simple cause–effect relation. Rather, traumatic events

may act as precipitants, the response to which is determined by individual vulnerabilities (6–8). These vulnerabilities may include biological determinants (1,9), cumulative life stressors (2,10), previous mental health problems, and a family history of mental illness (11,12). In addition to individual strengths and vulnerabilities, external factors such as social supports and the recovery environment have been found to influence responses to traumatic events (13–15). An early assumption often made in the social-support literature was that interactions between individuals in a social network are basically positive and that the existence of a support network will reduce traumatic distress (16). However, social-support studies of persons facing stressful situations have often overlooked the potentially troublesome aspects of interpersonal relationships and the uncertain benefits of so cial supports (17). Negative relationships may actually explain more of the variance in adjustment levels than do positive relationships (18). Further, while a positive socialsupport network may be associated with reduced distress, a cause–effect association between these variables cannot be inferred.

Manuscript received January 2000, revised, and accepted August 2000. 1Associate Professor, Institute for Medical Sciences and Faculty of Social Work; Director, Centre for Applied Social Research, University of Toronto, Toronto, Ontario. 2Lecturer, School of Business and Economics, Wilfrid Laurier University, Waterloo, Ontario. 3Firefighter, Mississauga Fire and Emergency Services, Mississauga, Ontario. Address for correspondence: Dr C Regehr, Centre for Applied Social Research, 246 Bloor Street West, Toronto, ON M5S 1A1 e-mail: [email protected]

Can J Psychiatry, Vol 46, March 2001

Underlying personality factors may explain findings that associate higher levels of adjustment with higher levels of social support; one possible factor is an individual’s relational capacity, or ability to develop and sustain interpersonal relationships. From this perspective, attachment experiences, especially in childhood, become incorporated into perceptions of self and other. These mental templates for significant relationship patterns not only determine future interactions but 156

March 2001

Individual Predictors of Posttraumatic Distress: A Structural Equation Model

also are likely to be self-confirming (19). When early relationships with caregivers are marked by hostility, victimization, and blaming, later relationships are bound to be seen through the filter of these past experiences (20). In this regard, social support is increasingly being viewed as a process of cognitive appraisal or a property of the person, rather than as an actual reflection of the transactions between individuals in a particular situation (17). That is, individuals with limited relational capacity may posses cognitive structures or schemas that emphasize negative factors and, consequently, underestimate the amount of support they receive. Further, the nature of one’s individual expectations and relationship skills plays a large part in determining both the types of people with whom one will associate and how they will be accepted (21). Thus, the ability to garner and use social support is in itself a strength and an effective coping resource. This research seeks to model the interrelation of relational capacity, perceptions of social supports, and posttraumatic distress. In this model, relational capacity is determined by 3 subscales of the Bell Object Relations and Reality Testing Inventory (BORRTI) that measure sense of alienation, insecurity, and social incompetence (22). Perceived support is determined by perceptions of support received from friends, family, and spouse or partner. Posttraumatic distress is measured by the 2 subscales of the Impact of Events Scale (IES) (23), Intrusion and Avoidance, and the Beck Depression Inventory (BDI) (24). We hypothesized that relational capacity contributes to perceived social support, which in turn contributes to the level of posttraumatic distress. Method This research was conducted within the Metropolitan Fire Brigade (MFB) of Melbourne and the Country Fire Authority (CFA) of Victoria, Australia. A group of 164 firefighters completed questionnaires administered by the researchers; of these, 99 were professional firefighters from the MFB and 65 were volunteer firefighters from the CFA. The mean age of respondents was 37.5 years, with an age range of 16 to 63 years. Of the respondents, 78% were married or living common-law; only 3.5% were divorced or separated. The remaining 18% were single. Firefighters in the sample had been with the fire service from 1 month to 38 years, with a mean of 12 years. Of the sample, 127 (78%) indicated that they had been exposed to at least 1 critical incident at work, including death of a colleague, injury on duty, mass casualties, or the death of a child. Respondents with the rank of officer were more likely to report having experienced a critical incident (P = 0.005). In addition, as years of employment increased, so did the incidence of reporting a critical incident (P = 0.001). Measures Surveys. In contrast to other studies, which mailed out questionnaires (25), surveys were administered and collected by

157

the researchers. Small groups (n = 5 to 20) completed questionnaires at an outstation or central fire hall. All participants were presented with information describing the study and its risks and benefits. Subsequently, each participant provided written consent. This procedure was approved by the university ethics committee. Demographic Data. Demographic variables were collected via a questionnaire that covered such items as age, sex, marital status, education, years in emergency-service work, perceived social support, and exposure to critical incidents. Current Level of Distress. This variable was measured by the BDI and the IES. The BDI is a self-report scale that assesses the presence and severity of affective, cognitive, motivational, vegetative, and psychomotor components of depression (24). Initially standardized on 606 psychiatric inpatients and outpatients, the reported reliability coefficient is 0.86. Test–retest reliabilities are 0.48 for psychiatric patients after 3 weeks and 0.74 for undergraduate students after 3 months. The BDI is now one of the most widely used measures of depression, both in clinical practice and with nonclinical research populations. The IES (23) assesses the experience of posttraumatic stress for any specific life event. It extracts dimensions that parallel the defining characteristics of DSM-IV PTSD: signs and symptoms of intrusive cognitions and affects occurring together or oscillating with periods of avoidance, denial, or blocking of thoughts and images. The 2 subscales, Intrusion and Avoidance, have high internal consistency, with Cronbach’s alphas of 0.78 and 0.82. Test–retest reliability is 0.87. A score of 26 or more is considered indicative of a diagnosis of PTSD (10). Relational Capacity. The BORRTI is a 90-item self-report that measures patterns of interpersonal relating (26). The instrument was tested on 934 subjects, including individuals who sat on the board of directors of a social agency and a business organization, undergraduate students without identified pathology, psychiatric outpatients, and psychiatric inpatients. The internal consistency reliability estimate of the subscales (Cronbach’s alpha coefficient) ranges from 0.90 to 0.78. On the basis of this testing, Bell reports nonpathological norms for all subscales and provides an interpretation of personality profiles associated with elevations of individual subscales and combined subscales (22). The Model We used structural equation modelling to develop a framework for understanding traumatic reactions. This framework incorporates both social supports and patterns of interpersonal relating. The model evaluated in this study (Figure 1) comprises 3 constructs (latent variables) and 9 measurement variables. The construct “relational capacity” was treated as

158

The Canadian Journal of Psychiatry

Vol 46, No 2

Figure 1. PARTNER ALIENTATION

1.00 1.00

FRIENDS -1.40

0.32

INSECURITY 1.51

RELATIONAL CAPACITY

0.81 SOCIAL INCOMPENTENCE

0.94

SUPPORT -0.45

DISTRESS

0.82 1.00

FAMILY

INTRUSION

0.96 1.31

0.30 AVOIDANCE

DEPRESSION

Chi-Square=26.09, df=23, P-value=0.29670, RMSEA=0.037

exogenous, the construct “support” as intermediate endogenous, and the construct “distress” as endogenous. Relational capacity was estimated using 3 subscales of the BORRTI (26): “alienation,” “insecurity,” and “social incompetence.” Support was estimated using 3 Likert scales that were included in the questionnaire asking individuals to rate the degree of support they received from their partner or spouse, their friends, and their family. Distress was estimated using the 2 subscales of the IES (23), Intrusion and Avoidance, and the BDI (24). When using structural equation modelling, we assume that the relations are linear and that the underlying measurement and latent variable are continuous. Since the 3 variables (perceived support from partner, friends, and family) were measured using 5-point Likert scales, the polychoric was used to calculate the correlation between these ordinal variables, and the polyserial correlation was used to calculate the correlation between the same ordinal variables and continuous variables measuring relational capacity and distress (27–29). Joreskog and Sorbom’s Prelis (30) was used to generate the correlation and asymptotic covariance matrices. Lisrel 8.3, a structural equation modelling program (30), was used to test the model’s fit. We employed the Weighted Least Squares (WLS) estimation technique, which is appropriate for analyzing the underlying parameters, using a polychoric–polyserial correlation matrix (30). The WLS estimation technique is also desirable because it is an asymptotically distribution-free method and does not require normality in the variables (31). Results Scores on the IES indicated that firefighters experience high degrees of traumatic stress reactions. Eleven firefighters

(7%) had scores indicating that they met DSM-IV criteria for PTSD. A further 68% had scores indicating moderate-level symptoms. Thus, only one-quarter of the sample scored in the mild- or no-symptom range. On the BDI, 3% had scores indicating that they were suffering from severe depression, and 19.5% were suffering from moderate levels of depression. Length of time since the most recent traumatic event encountered was not significantly associated with BDI or IES scores. The number of critical events experienced, however, was positively correlated with both the IES (r = 0.24, P = 0.001) and the BDI (r = 0.18, P = 0.05). Means on the subscales of the BORRTI did not differ significantly between those who reported encountering a critical incident and those who did not. The Model The overall fit of the model was excellent (χ2 = 26.09, df 23, P = 0.30), which indicates that the hypothesized model did not differ significantly from the data. The “root mean square error of approximation” (RMSEA) is a measure of discrepancy per degree of freedom between the model and the population (29). A RMSEA of 0.05 or lower indicates a close fit. The model had a RMSEA = 0.037. Each of the additional fit statistics likewise indicated an excellent model fit, Comparative Fit Index (CFI) = 1.00, Goodness of Fit Index (GFI) = 0.99, Adjusted GFI = 0.98, Normed Fit Index (NFI) = 0.97. Thus, the hypothesized model fit the data very well. Relational Capacity To set the metric of the latent variable “relational capacity,” the loading of the first measurement variable “alienation” was set to 1 (λx11 = 1). The loadings of “insecurity” (λx21 = 1.51, P < 0.001) and “social incompetence” (λ31 = 0.81, P < 0.001) were both large and significant. “Relational capacity”

March 2001

Individual Predictors of Posttraumatic Distress: A Structural Equation Model

Table 1. Factor loadings, z-scores and P-values for the posttraumatic distress model Measure and variable

Factor loading

z score

P value

Relational capacity Alienation

1.00a

—

—

Insecurity

1.51

z = 7.29

P < 0.001

Social incompetence

0.81

z = 9.83

P < 0.001

Perceived support Partner

1.00a

Friends

0.94

z = 7.30

P < 0.001

Family

0.82

z = 6.96

P < 0.001

1.00a

—

—

Distress Intrusion Avoidance

0.96

z = 7.04

P < 0.001

Depression

1.31

z = 6.34

P < 0.001

Relational capacity–support

–1.40

z = –4.01

P < 0.001

Support–distress

–0.45

z = –3.60

P < 0.001

had a significant direct negative effect on “support” (γ11 = –1.40, t = – 4 .01, P < 0.001), indicating that as disturbances in relational capacity increased, the amount of support that an in di vid ual per ceived that they re ceived from oth ers decreased. Support As with “relational capacity,” the metric of the latent variable “support” was set to that of the first measurement variable, “perceived support from spouse or partner,” by fixing the loading to 1 (λ11 = 1). The loadings of “perceived support from friends” (λy21 = 0.94, P < 0.001) and “perceived support from family” (λy31 = 0.82, P < 0.001) were both large and significant. “Support” had a significant negative direct effect on “distress” (β21 = – 0 .45, t = –3.60, P < 0.001), indicating that as perceived support increases, individual distress decreases. Distress As with “relational capacity,” the metric of the latent variable “distress” was also set to that of the first measurement variable, the Intrusion subscale of the IES, by fixing the loading to 1 (λy42 = 1). The loadings of the Avoidance subscale of the IES (λy52 = 0.96, P < 0.001) and depression as measured by the BDI (λy62 = 0.82, P < 0.001) were both large and significant. Thus, all the measurement variables used in the study had significant loadings on their associated constructs (latent variables) and were relatively large. All structural parameters were also statistically significant, indicating that relational capacity does have a significant negative relation to support, and in turn, support has a significant negative relation to distress. The squared multiple correlations (R2) for the structural equations were as follows: for the support construct R2 = 0.58,

159

and for the distress construct R2 = 0.88. This indicates that 58% of the variance in perceived support was accounted for by relational capacity, and 88% of the variance in levels of distress could be accounted for by the combined effects of relational capacity and perceived support. Discussion As the dose-effect model for understanding response to trauma becomes increasingly suspect, researchers are considering other factors that may account for variations in traumatic symptomatology. Several authors have suggested that social support is an important variable in determining traumatic responses (13,14). Others have pointed to the centrality of individual personality factors in influencing the continuation of posttraumatic distress (6,8). This study determined the interrelation of relational capacity (a component of personality), perceptions of social support, and posttraumatic distress. Based on the previous work of the first author (32), it was postulated that, of the multiple personality factors influencing behaviour and response to traumatic events, relational capacity (the ability to sustain interpersonal relationships) is a key individual factor contributing to persistent posttraumatic stress and depressive symptoms. This study further hypothesized that relational capacity contributes to the development of social supports and to perceptions of support from key members of an individual’s network. Finally, it was hypothesized that social supports would influence levels of posttraumatic stress and depression in a sample of firefighters exposed to critical events. A structural equation model of these hypothesized relations strongly supported the theoretical model. In this study, therefore, individuals whose BORRTI scores indicated a basic lack of trust in relationships, sensitivity to rejection, a tendency to be easily hurt by others, and difficulty in making friends also perceived less support from those in their social network following a traumatic event. Further, these individuals scored significantly higher on measures of depression and posttraumatic distress. Indeed, the combined effects of relational capacity and perceived social support accounted for 88% of the variance in distress among firefighters exposed to critical events. The amount of time elapsed since the event was not significantly associated with level of dis tress, suggesting that the impact of individual variables does not diminish with time. These findings support the notion that individuals encountering traumatic events will vary significantly in their levels of distress, based on their ability to develop and sus tain sup por tive re la tion ships and their perceptions of the responses of others. In addition, however, a mild positive relation was found between the number of critical events encountered and both measures of traumatic distress, which supports the contention that individual variables alone do not determine level of distress and that the dose of

160

The Canadian Journal of Psychiatry

trauma is indeed a factor. This would be expected to be more significant in populations wherein the traumatic event was more personal and of greater intensity. Study Limitations In this cross-sectional study, subjects were queried simultaneously about past traumatic events and about their current level of psychological functioning. Thus, it was hypothesized that relational capacity contributes to perceptions of social support, which in turn contribute to levels of distress. Briere, however, cautions that the correlational and retrospective nature of many trauma studies can blur cause and effect (33); it could therefore be argued that, although relational capacity and social support may mediate traumatic response, the reverse could also be true. That is, time elapsed and current level of distress may influence perceptions of relationships. While it is not possible to compare the level of social support perceived by individuals who encountered critical events with that perceived by those who did not, there were no significant differences on measures of relational capacity between individuals who reported encountering critical events and those who did not. This suggests that relational capacity was not influenced by exposure to traumatic stimuli. However, perceived social support and relational capacity may have been influenced by current level of distress. Conclusion The results of this exploratory study support the hypothesis that relational capacity is a significant factor in explaining perceptions of social support in firefighters exposed to critical events and that perceived social support influences levels of depression and posttraumatic symptoms. This suggests that, while some emotional response to disturbing events may be normal, the severity of symptoms covaries with an individual’s ability to develop and sustain supportive relationships to buffer the impact of events. Individuals who are unable to trust others, are sensitive to rejection, are easily hurt by others, and experience difficulty making friends are more likely to experience higher levels of distress following a critical event. The implications of these findings must be approached with caution. Among its benefits, the concept of PTSD suggests that traumatic response is not pathological; thus, it removes the burden of victim-blaming from the individual. While relational capacity may influence perceptions of social support and traumatic distress, it must be acknowledged that previous life challenges and opportunities have influenced the ability to develop and sustain interpersonal relationships in individuals encountering traumatic events. From this perspective, these findings can assist in the development of interventive programs that will address underlying patterns of interpersonal relating and cognitive appraisals of others.

Vol 46, No 2

Clinical Implications • Relational capacity is a significant factor in explaining perceptions of social support following a traumatic event.

• Perceived levels of social support influence levels of depression and posttraumatic distress experienced after a traumatic event.

• Intervention programs for trauma must be prepared to address underlying patterns of interpersonal relating as well as immediate symptoms of distress.

Limitations • Current levels of distress may influence relational capacity and perceptions of social supports.

• The cross-sectional design of this study does not allow for measurement of change in relational capacity or perceived social support.

• Generalizability is limited because participants were firefighters who encountered traumatic events on the job.

Acknowledgement This research was generously supported by an institutional grant from the Social Sciences and Humanities Research Council of Canada.

References 1. Yehuda R, McFarlane A. Conflict between current knowledge of post-traumatic stress dis or der and its origi nal con cep tual ba sis. Am J Psy chia try 1995;152:1705–13. 2. Mollica R, McInnes K, Poole C, Tor S. Dose-effect relationships of trauma to symptoms of depression and post-traumatic stress disorder among Cambodian survivors of mass violence. Br J Psychiatry 1998;173:482–8. 3. Resnick H, Kilpatrick D, Best C, Kramer T. Vulnerability-stress factors in development of posttraumatic stress disorder. J Nerv Ment Dis 1992;180:424 –30. 4. Marmar C, Weiss D, Metzler T, Delucchi K, Best S, Wentworth K. Longitudinal course and predictors of continuing distress following critical incident exposure in emergency services personnel. Journal of Nervous and Mental Disordors 1999;187:15–22. 5. Bowman M. individual differences in posttraumatic distress: problems with the DSM-IV model. Can J Psychiatry 1999;44:21–33. 6. Paris J. Nature and nurture in psychiatry: a predisposition stress model of mental disorders. Washington (DC): American Psychiatric Press; 1999. 7. Yehuda R. Biological factors associated with susceptibility to posttraumatic stress disorder. Can J Psychiatry 1999;44:34 –9. 8. van der Kolk B, McFarlane A, Weisaeth L. Traumatic stress: the effects of overwhelming experience on mind, body and society. New York: Guilford Press; 1996. 9. True W, Rice J, Eisen S, Heath A, Goldberg J, Lyons M, Nowak J. A twin study of genetic and environmental contributions to liability for posttraumatic stress symptoms. Arch Gen Psychiatry 1993;50:257. 10. McFarlane A. The longitudinal course of posttraumatic morbidity. J Nerv Ment Dis 1988;176:30–9. 11. McFarlane A. The aetiology of post-traumatic stress disorder following a natural disaster. Br J Psychiatry 1988;152:116–21. 12. Skodol A, Schwartz S, Dohrenwend B, Levav I, Shrout P, Reiff M. PTSD symptoms and comorbid mental disorders in Israeli war veterans. Br J Psychiatry 1996;169:717–25. 13. Leffler C, Dembert M. Posttraumatic stress symptoms among US Navy divers recov er ing TWA Flight 800. Jour nal of Nerv ous and Men tal Dis or ders 1998;186:574–7. 14. King L, King D, Fairbank J, Keane T, Adams G. Resilience-recovery factors in post-traumatic stress disorder among female and male Vietnam veterans: hardiness, postwar social support, and additional stressful life events. Journal of Personality and Social Psychology 1997;74:420–34. 15. Weiss D, Marmar C, Metzler T, Ronfeldt H. Predicting symptomatic distress in emergency services personnel. J Consult Clin Psychol 1995;63:361–8. 16. Lyons J. Strategies for assessing potential for positive adjustment following trauma. Journal of Traumatic Stress 1991;4:93–111. 17. Coyne J, DeLongis A. Going beyond social support: the role of social relationships in adaption. J Consult Clin Psychol 1986;54:454–60. 18. Wortman C. Social support and the cancer patient: conceptual and methodological issues. Cancer 1984;53:2339–60.

March 2001

Individual Predictors of Posttraumatic Distress: A Structural Equation Model

19. Hadley J, Holloway E, Mallinckrodt B. Common aspects of object relations and self-representations in offspring from disparate dysfunctional families. Journal of Counselling Psychology 1993;40:348–56. 20. van der Kolk B, Hostetler, Herron N, Fisler R. Trauma and the development of border line per son al ity dis or der. Psy chi at ric Clin ics of North Amer ica 1994;17:715–30. 21. Bowlby, J. The making and breaking of affectional bonds. London: Tavistock Publications; 1979. 22. Bell M. An introduction to the Bell object relations and reality testing inventory. Los Angeles: Western Psychological Services; 1991. 23. Zilberg N, Weiss D, Horowitz M. Impact of event scale: a cross-validation study and some empirical evidence supporting a conceptual model of stress response syndromes. J Consult Clin Psychol 1982;50:407–14. 24. Beck A, Beamesderfer A. Assessment of depression: the depression inventory. In: Pichot P, editor. Psychological measurements in psychopharmacology. Paris: Karger Basel; 1974. p 151–69.

161

25. Bryant R, Harvey A. Posttraumatic stress reactions in volunteer firefighters. Journal of Traumatic Stress 1996;9:51–62. 26. Bell M, Billington R, Becker B. Scale for the assessment of reality testing: reliability, validity and factorial invariance. J Consult Clin Psychol 1985;53:506–11. 27. Bollen K. Structural equations with latent variables. New York: John Wiley and Sons; 1989. 28. Muthen B. A general structural equation model with dichotomous, ordered categorical and continuous latent variable indicators. Psychometrika 1984;49:115–32. 29. Jorsekog K, Sorbom D. Lisrel 8: user’s reference guide. Chicago: Scientific Software International; 1999. 30. Jorsekog K, Sorbom D. Prelis 2: user’s reference guide. Chicago: Scientific Software International; 1999. 31. Browne M. Asymptotic distribution free methods in analysis of covariance structures. British Journal of Mathematical and Statistical Psychology 1984;37:62–83. 32. Regehr C, Marziali E. Recovery from sexual assault: a relational perspective. Journal of Nervous and Mental Disorders 1999;187:618–23. 33. Briere J. Methodological issues in the study of sexual abuse effects. J Consult Clin Psychol 1992;60:196–203.

Résumé— Prédicteurs individuels du stress post-traumatique : un modèle d’équation structurelle Objectif : La recherche récente a remis en question le modèle « dose-effet » pour comprendre la réaction au traumatisme, et s’est plutôt penchée sur la contribution des facteurs de la personnalité et de l’environnement. Cette étude cherche à modeler l’interrelation de la capacité relationnelle (une composante de la personnalité), des perceptions du soutien social et de la détresse posttraumatique. Méthode : Un groupe de pompiers (n = 164) ont répondu à des questionnaires portant sur l’exposition à des événements traumatiques, le soutien social, le niveau de détresse actuel et la capacité relationnelle. L’établissement d’un modèle d’équation structurelle a servi à élaborer un cadre de compréhension des réactions traumatiques. Résultats : En général, le modèle hypothétique cadrait admirablement bien. La capacité relationnelle avait un effet négatif significatif sur le soutien, ce qui indique que le soutien social perçu diminuait à mesure qu’augmentaient les perturbations de la capacité relationnelle. Le soutien social perçu avait un effet négatif significatif sur le niveau de détresse. Conclusion : Même si une certaine réaction émotive à des événements bouleversants est normale, la gravité des symptômes varie avec la capacité de la personne de nouer et d’entretenir des relations de soutien pour absorber l’effet des événements.