BEREAVEMENT BY TRAUMATIC MEANS: THE COMPLEX SYNERGY OF TRAUMA AND GRIEF

Abstract

In the universe of life events, traumatic loss and traumatic stress intersect a great deal, both in event dimensions and psychological impact. Typically, loss by traumatic means (e.g., homicide) is conceptualized as a traumatic stressor event that can lead to posttraumatic stress disorder (PTSD; as defined in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders [DSMIV]). However, grief is a distinct individual, social, and relational experience. Until recently, bereavement-related biopsychosocial phenomena were not taken into account in conceptualizations of loss by traumatic means. Traditionally, the study of adaptation to trauma and the study of adaptation to loss have been distinct, with rare exceptions (e.g., Eth & Pynoos, 1994; Green et al., 2001). The boundaries between traumatic stress and PTSD, complicated or chronic bereavement as a mental health outcome independent of the nature of the loss, and traumatic bereavement (i.e., loss by traumatic means) and traumatic grief (the unique mixture of trauma and loss) have not been examined sufficiently. While the phenomenologies of PTSD and complicated or traumatic grief have much in common, the study of these “syndromes” rarely includes comprehensive evaluations of both outcomes. For example, while most of the research on the aftermath of September 11 studied extensively PTSD and depression (e.g., Galea et al., 2002; Silver et al., 2002), no study has examined loss reactions, or the interplay between trauma and loss and their outcomes.

We review the complex interplay between trauma and loss by traumatic means. First, we address the distinctions between normal and traumatic loss, and complicated and traumatic grief, and contrast these experiences with PTSD. Second, we discuss the role of various mediators such as concurrent or life-span trauma exposure and interpersonal factors, particularly the degree of attachment to the individual or group traumatically lost. We also offer a more integrated and focused view of traumatic grief, its predictors, and future directions for the integrative study of trauma and loss outcomes.

Trauma and Loss

Loss reactions are universal yet tremendously variable and, regardless of the circumstances of the loss, grief is typically relatively short-lived. Painful as the experience can be, most people accommodate loss and regain normal functioning (e.g., Lindemann, 1944; Parkes & Weiss, 1983). The form and course of bereavement are highly individualized. As a result, it has been difficult to generate a boundary between normal and abnormal grief. Yet, clearly, there is a small percentage of individuals who fail to return to normal functioning; they are stuck with a degree of steady mourning and functional impairment. Traditionally, loss has been conceptualized psychodynamically, with the assumption that when the “normal resolution” of the emotional toll of loss is avoided or blocked, the necessary “grief work” (i.e., the “working-through” process) is thwarted and grief becomes chronic, enduring, and disabling. This view has been challenged by a number of recent studies showing that individuals who suppress their emotional reactions to a loss or fail to exhibit outpourings of grief do not fare worse over time (e.g., Stroebe, Stroebe, Schut, Zech, & van den Bout, 2001).

The study of difficult bereavement has a long history in psychiatry and psychology. Starting with the Lindemann (1944) landmark study of acute grief in survivors of a horrific nightclub fire, many of whom were both bereaved and directly traumatized by the disaster, and followed by Bowlby's work on separation loss, which is considered a developmental trauma (Bowlby, 1969), the study of unresolved, complicated reactions to loss has aimed to appreciate the phenomenology and to define and predict chronic, unresolved grief reactions.

Green (2000) and Green et al. (2001) argued cogently that loss by traumatic means should be treated as a traumatic stressor, and that the resulting chronic condition that arises in a small percentage of cases should be classified as PTSD. In this conceptualization, violent and unexpected loss results in severe feelings of personal vulnerability and forces the individual to confront the prospect of death, creating intense anxiety, which arguably is the psychological aftereffect common to all traumatic stressors.

Consistent with this framework, Green et al. (2001) examined the effects of traumatic loss in comparison to non-loss-related trauma. These researchers studied individuals with a single traumatic bereavement, individuals with a single non-loss-related trauma, and individuals with no traumatic experience. They found that 16% of the persons who suffered loss by traumatic means met the criteria for PTSD, and 22% of them also met lifetime PTSD criteria. The prevalence of major depression was no higher in the traumatic loss group than in the other two groups, supporting the hypothesis that a“post-loss syndrome” is not simply depression. The most stigmatized deaths and those associated with intent tended to produce higher rates of stress disorder. In fact, loss by traumatic means led to more severe intrusive symptoms and greater functional impairment in comparison to a group of individuals who suffered physical assault, which suggests that loss by traumatic means may be more pernicious than direct trauma. Unfortunately, Green et al. (2001) failed to take into account the nature and extent of the attachment relationship in those who lost loved ones to violence. In addition, they failed to directly contrast PTSD as an outcome variable with symptoms of chronic grief.

We argue that the symptoms of PTSD fail to sufficiently capture the unique experiences of those who suffer from chronic grief as a result of violent loss of an important attachment figure. While there is no doubt that loss of an important attachment figure by violent means is potentially traumatizing and could result in symptoms of PTSD, there is also sufficient empirical evidence and compelling alternative conceptual frameworks to argue against a restrictive and narrow conceptualization of loss by traumatic means as a psychological trauma.

The “loss as trauma” framework (e.g., Green, 2000) fails to sufficiently acknowledge the unique biological, psychological, and social behavior implications of bereavement, which affects adaptation to loss by violence. In addition, within the field of traumatic stress, there is general consensus that certain types of traumatic events in certain contexts or developmental periods lead to unique posttraumatic outcomes. For example, although interpersonal trauma (e.g., incest, sexual assault, physical assault by caregivers and attachment figures) is defined in the same way as non-interpersonal trauma (e.g., motor vehicle accident) in the diagnostic framework, interpersonal trauma leads to a different repertoire of posttraumatic deficits and liabilities while sharing the same summary label of “PTSD” (e.g., Herman, 1992; Zlotnick, Zakriski, Shea, & Costello, 1996). These effects might be further mediated by the unique meaning attributed to interpersonal trauma, which is experienced as abetrayal of attachment (e.g., Freyd, DePrince, & Zurbriggen, 2001). Thus, traumatic experiences have a different psychological impact depending on the meaning attributed to the event. Within this framework, the psychological and psychiatric impact of loss by traumatic means is mediated by how the individual construes the implications of the event.

The Concept of Complicated Grief

Horowitz et al. (1997) were the first to propose a complicated grief disorder to be included in the diagnostic nosology. The symptoms proposed were (a) unbidden memories of intrusive fantasies related to the lost relationship; (b) strong spells or pangs of severe emotion related to the lost relationship; (c) distressingly strong yearnings or wishes that the deceased were there; (d) feelings of being far too much alone or personally empty; (e) excessive avoidance of people, places, or activities that remind the person of the deceased; (f) unusual levels of sleep interference; and (g) loss of interest in work, social, caretaking, or recreational activities to a maladaptive degree. In a small-scale exploratory study, Horowitz et al. (1997) found that 41% of subjects studied 6 months postloss received their diagnosis of complicated grief at14 months. The researchers failed, however, to disaggregate the unique consequences of loss by traumatic means. Their proposed diagnostic category was intended to capture chronic problems arising from any type of bereavement experience. In a subsequent article, they hypothesized that high anxiety (as an individual difference characteristic) can lead to cognitive avoidance of processing distressing grief-related information, which, in turn, can lead to more severe and chronic grief reactions (Horowitz, Bonanno, & Holen,1993). Given the prevalence estimates from their pilot study, it appears to us that the diagnostic criteria offered for complicated grief, and the decision rules used to define caseness, are overly pathologizing a normal range of chronic postloss adaptation.

Traumatic Grief

Recently, Prigerson and colleagues have proposed the construct of traumatic grief as a distinctive psychopathological condition stemming from chronic bereavement. Relying primarily on a large sample of elderly widows and widowers, Prigerson and others developed diagnostic criteria for traumatic grief syndrome, a pathological response to a loss of a significant other that is distinct from depression, anxiety, and PTSD (Prigerson et al., 1999).

The diagnostic criteria for traumatic grief center on two components: the separation distress of losing an attachment figure and the traumatic distress of adjusting to life without that figure (Prigerson et al., 1999). Correspondingly, traumatic grief symptoms include some intense, impairing grief symptoms (e.g., yearning for the lost person, loneliness) as well as symptoms of PTSD (e.g., intrusive, distressing experiences of the loss experience; Prigerson & Jacobs, 2001; Prigerson et al., 1999). The symptoms of separation distress are (a) intrusive, distressing preoccupation with the deceased; (b) yearning, longing, and pining; (c) searching for the deceased; and (d) extreme loneliness. The symptoms of traumatic distress include (a) feeling unfulfilled without the deceased; (b) avoidance of painful reminders of the loss; (c) futility about the future; (d) feeling that a part of the self has died; (e) numbness and detachment; (f) shattered world view (regarding trust, security, control); (g) feeling shocked, stunned, and dazed; (h) disbelief about the death; (i) emptiness; (j) taking on symptoms or harmful behaviors of the deceased; and (k) bitterness. Since the various symptoms of traumatic grief are normally present acutely, the syndrome is chiefly defined by 6 or more months' persistence of the problems and the degree of enduring functional impairment in individuals complaining of difficulty recovering from loss. At this early stage of construct development, it is unclear whether traumatic grief should be considered only in the case of loss by traumatic means, or more broadly in the case of any significant attachment loss.

Traumatic grief is associated with considerable psychiatric and physical health morbidity, such as high blood pressure (Prigerson et al., 1997, 2001), cancer (Prigerson et al., 1997), cardiac events (Prigerson et al., 1997), ulcerative colitis (Lindemann, 1944), suicidality (Prigerson et al., 1997), and global dysfunction (Prigerson et al., 1997). Importantly, there are severalindi-cations that grief reactions, unlike depression, are not effectively treated with antidepressants or interpersonal psychotherapy (Pasternak et al., 1991; Reyndols et al., 1999). These findings suggest the possibility of a major public health problem in individuals who have suffered traumatic loss of significant attachment figures.

Some have argued that the term “traumatic bereavement” should be employed to describe the unique experience of losing a significant other due to sudden, violent, or accidental means (Raphael & Martinek, 1997; Stroebe, Schut, & Finkenauer, 2001). Survivors who suffer through a traumatic bereavement have to cope with the trauma and any resulting stress in addition to the death and the grieving process (Raphael & Martinek, 1997). Having to deal with posttraumatic stress as a result of a traumatic loss can interfere with the grieving process, leading to postloss functional impairment. Raphael and Martinek (1997) suggest that traumatic bereavement is associated with more adverse health outcomes than normal, uncomplicated bereavement; however, there is little sound, empirical research on this topic. The data that do exist suggest that mental health outcomes of traumatic bereavement follow a longer course, are more adverse, and feature both posttraumatic stress and grief phenomenology (Raphael & Martinek, 1997). Nevertheless, although the construct of traumatic grief proposed and studied by Prigerson and colleagues can arise purportedly from any loss, it may well represent a unique class of symptoms that best describe the experiences of individuals bereaved by traumatic means (Prigerson et al., 1999b; Prigerson & Jacobs, 2001; Jacobs, Mazure, & Prigerson, 2000). However, this is an empirical question that has yet to be addressed in the literature.

Traumatic Grief From Vıolence and Mass Casualty Events

After the 1995 Oklahoma City bombing, which, prior to 9/11, was the worst terrorist attack on American soil, the immediate psychological needs of survivors were studied extensively by Pfefferbaum and her colleagues (Tucker, Dickson, Pfefferbaum, McDonald, & Allen, 1997; Pfefferbaum et al., 2001). A convenience sample of adults exposed to the attack on the Alfred P. Murrah Federal Building reported experiencing symptoms of PTSD 6 months later (Tucker et al., 1997). These adults differed in the amount of exposure they had to the bombing; some had lost someone close to them, some were injured themselves, and others had been exposed primarily through media outlets. In an effort to determine the number of people who suffer from traumatic grief following a terrorist attack, Pfefferbaum and colleagues (2001) reexamined a subset of individuals who were bereaved after the Oklahoma City bombing. This subset was composed of 50% of their original study group. The researchers found that the victims'grief response accounted for a significant portion of the variance in victims' posttraumatic stress symptoms. More specifically, victims' scores on the Texas Inventory of Grief covaried with their self-reported levels of PTSD symptoms. When taken together, grief scores and PTSD symptomatology combined to predict current levels of functional impairment (Pfefferbaum et al., 2001). This research underscores the phenomenological synergy of trauma and grief and the combined impact on post-event functional impairment. However, the study of grief from the Oklahoma City bombing is an insufficient test of the construct of traumatic grief, because symptoms were not directly evaluated. In addition, very few people in the study group lost a significant attachment figure (i.e., close family member; 10% of sample). The overwhelming majority of victims lost acquaintances (Pfefferbaum et al., 2001). Thus, the external generalizability of the study results is limited, particularly in reference to the aftermath of mass casualties on September 11, 2001.

Combat in the war zone has provided a laboratory to study the effects of traumatic bereavement on human functioning over the life span. Exposure to violence-related trauma and the loss of close comrades and friends in battle and war captivity accounts for variance in postwar distress and social dysfunction (e.g., Holman & Silver, 1998; Neria, 2001; Neria et al., 2000). Green, Grace, Lindy, Gleser, and Leonard (1990) evaluated a large group of Vietnam war veterans and found that 70% of those with PTSD, as compared to 29% of those without, reported the loss of a “buddy,” generally suggesting that the dual burden of loss by traumatic means and direct trauma results in worse chronic adaptation to war. In a war zone, soldiers who lose comrades are typically unable to grieve because of concurrent demands; very often, they cannot benefit from funeral rituals (e.g., bodies may have not been retrievable or the severely wounded were evacuated); and their coping resources are generally drained managing ongoing threat and anxiety.

A clinical study by Eth and Pynoos (1994) revealed that children who were both traumatized and bereaved by witnessing their parents violently murdered hadacute posttraumatic stress reactions, which interfered with their ability to successfully grieve. Additionally, these children-survivors tended to regress developmentally, leading to impaired school performance and an inability to trust others and to form meaningful attachments. Similarly, children who lost their fathers in the war in Bosnia were shown to suffer from depression as well as reactions to trauma (Zvizdic & Butollo, 2001). In a survey study, sons and daughters of fathers who disappeared or were killed in the war had severe difficulties adjusting to their postwar world when compared to a group of children who had been separated from their fathers but were reunited after the war (Zvizdic & Butollo, 2001). Notably, children whose fathers had disappeared had the most severe symptoms of depression and posttraumatic stress. The researchers believe that the elevated levels of symptoms in this group were due to the ambiguity and uncertainty surrounding the fathers' disappearance, which added yet another psychological stressor to the children's lives (Zvizdic & Butollo, 2001). Prigerson and colleagues describe another adverse effect of traumatic bereavement, finding that young adult friends of suicide victims who demonstrated high levels of traumatic grief also expressed high levels of suicidal ideation (Prigerson, Bridge et al., 1999).

What Factors Mediate Loss-Grief Relations?

The bereavement literature suggests a number of factors, independent of the nature of the loss (traumatic vs. nontraumatic), that influence the course and the outcome of bereavement. These factors include characteristics and events specific to the individual (intrapersonal), as well as relational variables between the individual and other people in the environment or the social network of the bereaved (interpersonal). Virtually no empirical studies have been conducted examining the mediators and moderators of responses to bereavement from traumatic means, although a handful of reports have suggested that experiencing past trauma or previous loss may complicate or prolong the bereavement process (Raphael, 1997; Sanders, 1993; Green, 2000). These life experiences may lead to psychopathological symptoms that may worsen following the loss of a significant other.

With regard to various sociodemographic characteristics, research has shown that bereaved women generally adjust better to loss than their male counterparts (Stroebe & Schut, 2001). Also, studies have shown that widowers have relatively higher rates of both mortality and depressive symptoms than widows (Stroebe & Stroebe,1983; Umberson, Wortman, & Kessler, 1992).

Studies examining the role of age have found that young adults who experience a loss tend to have higher rates of mortality and grief symptoms than older bereaved adults (Helsing & Szklo, 1981; Maddison & Walker, 1967). This effect may be moderated by the expectedness of the loss, as accidents are the leading killer of young adults, but cancer and heart disease are common causes of death among the elderly (Stroebe & Schut, 2001).

Little is known about the mediating role that certain world religions play in the course of bereavement (i.e., Judaism, Buddhism, Christianity, etc.). However, Levy and colleagues (1994) and Nolen-Hoeksema and Larson (1999) showed that religious bereaved individuals experienced fewer depressive symptoms over time than nonreligious grievers. The effect of religiosity may be mediated by the social support one receives through church attendance (Nolen-Hoeksema & Larson, 1999).

Coping Behaviors

Certain coping styles, such as rumination, are quite maladaptive following a loss (Nolen-Hoeksema, 2001; Holman & Silver, 1998). Additionally, emotion-focused coping has been shown to be more effective than problem-focused coping, yet both are needed to successfully deal with a loss (Stroebe & Schut, 1999). Coping style may moderate the effect of gender on grief, as men generally tend to use problem-focused coping, while women typically use both forms of coping to deal with loss (Stroebe & Schut, 2001).

Perception of control has been shown to play a moderating role in the effect of expectedness of the loss on grief (Stroebe et al., 1988). Stroebe et al. (1988) found that bereaved individuals who did not expect the loss had more depressive symptoms and somatic complaints than those for whom the loss was expected, but only when the bereaved had an external locus of control.

Interpersonal Mediators and Moderators of the Grief Response

Parkes (2002) identified survivor vulnerabilities that are risk factors for complicated bereavement, such as dependence on the lost person and a lack of self-esteem. Self-esteem, like perception of control, was shown to moderate the negative impact of sudden loss on the grieving process (Stroebe et al., 1987). Sudden losses led to poor well-being in bereaved persons with low self-esteem. Additionally, survivors' attachment styles may put them at risk for complicated bereavement (Parkes & Weiss, 1983; Bowlby, 1980). Anxious-ambivalent attachment styles are associated with chronic grief (Parkes & Weiss, 1983; Bonanno & Kaltman, 1999), but work needs to be done to determine the impact of avoidant attachment styles on the grief response (Bonanno & Kaltman, 2001; Stroebe & Schut, 2001).

The loss of a child, spouse, sibling, parent, close friend, colleague, or acquaintance each holds a different meaning for an individual in terms of the relationship that is lost (Stroebe & Schut, 2001). Studies have shown that the loss of a child leads to more intense and persistent grief and depression than the loss of a sibling, spouse, or parent (Cleiren, 1991; Leahy, 1992; Nolen-Hoeksema & Larson, 1999; Sanders, 1979). Similarly, the loss of a close friend or family member could predict more difficulty than the loss of a colleague or an acquaintance. However, more research needs to be done to substantiate the relationship between quality of attachment relationship and mental health outcomes of traumatic loss in particular.

Attachment theory informs bereavement research by virtue of the fact that grief comes from mourning the loss of an attachment relationship. This grief is intensified if the relationship that is lost is particularly significant in the life of the bereaved (Weiss, 2001). Losing a loved one who was an integral part of the bereaved individuals social network, support system, or identity, or was a cherished companion and confidant, causes more pain and confusion in this time of distress (Shaver & Tancredy, 2001).

The support of family and friends in a time of need is vital to a person's ability to cope. After the loss of a loved one, the importance of this support becomes twofold (Stroebe & Schut, 2001). First, as in any crisis, to have someone with whom the individual in crisis feels comfortable to work through his or her emotions is an invaluable resource. A network of people is necessary to ease the added responsibility of planning appropriate ceremonies for the deceased and to alleviate some of the tasks of daily life that become a burden in this troublesome time. Second, the loss of a partner or companion can leave the bereaved with significant emotional loneliness, during which time social supports become even more essential (Stroebe & Schut, 2001).

Conclusions

As can be seen, many of the risk factors for complicated bereavement have been understudied, particularly in the context of traumatic loss. No study to date has investigated comprehensively the risk indicators and, more important, the mechanisms of risk associated with chronic postloss impairment (as well as the resilience variables that decrease risk).

It is safe to conclude that unpredictable loss by malicious violence is one of the most pernicious human experiences, creating the greatest risk for chronic postloss difficulties (e.g., Rynearson & McCreery, 1993; Spooren, Henderick, & Jannes, 2000; Zvizdic & Butollo, 2001; Pfefferbaum et al., 2001). In this bereavement context, the demands and outcomes are represented by a synergy of psychological trauma and grief. The study of loss by traumatic means, and in particular the psychological and psychiatric sequelae implicated by loss due to malicious violence, is relatively new. At present, there is no single paradigmatic approach but rather several competing theories conceptualizing the causes of chronic grief implicated in bereavement by traumatic loss.

There is also insufficient information on the phenomenology, clinical symptoms, clinical needs, and risk factors associated with critical loss, coupled with the added burden of direct traumatization experiences. Although rarely studied empirically, it has been argued cogently that when loss by traumatic means is coupled with the disruption caused by direct traumatization, this dual burden is uniquely onerous (e.g., Raphael & Martinek, 1997). This dual burden of loss by traumatic means on top of direct traumatization is emblematic of disasters generally and the aftermath of terrorist attacks especially (Pfefferbaum et al., 2001), yet the combined psychosocial consequences have been understudied. Given traumatized individuals' need to regain a sense of safety, comfort, and protection, it is expected that individuals who have suffered direct trauma and lost attachment figures will suffer from more chronic PTSD than individuals who have not lost important parts of their support systems.

Finally, in the bereavement field, the nature and quality of the attachment relationship is one of the most important determinants of the psychological impact of any type of loss (e.g., Parkes, 2002; Shaver & Tancredy, 2001; Bonanno & Kaltman, 1999). For example, it is plausible to assume that among those who suffered losses as a result of September 11, 2001, those who had the closest attachment relationship with the lost individual are at the greatest risk for chronic, traumatic bereavement-related disturbances. While it is necessary to evaluate the familial connection and the apparent degree of intimacy inferred from the labels individuals use to describe the nature of the relationship to the deceased (e.g., boyfriend, partner), it is also critical to evaluate the meaning and the implication of the loss for the individual. The latter will provide especially useful information we can use to refine treatment models for traumatic grief as a result of mass violence.

Biographies

References