| Disasters, A Psychological Perspective
Margaret Gibbs, Ph.D. and Kim Montagnino, M.A.
Psychological research has shown that disasters can cause serious mental health consequences for victims. These consequences take the form of Posttraumatic Stress Disorder and a variety of other disorders and symptoms which have been less investigated. The more stress, defined in a variety of ways, within the disaster, the more likely there are to be emotional consequences. Vulnerability factors within the victim operate in complex ways, but seem related to the extent of stress experienced by the victim and the available resources, broadly defined, with which to deal with it. The mental health profession has developed a variety of strategies with which to ameliorate the effect of disaster. Although recent research on single session debriefing has produced disappointing results, many techniques and therapies have been validated as successful interventions for disaster victims.
Unlike other disciplines, which have come more recently to the study of disasters, psychology has concerned itself with disasters' impacts on victims for much of its own short history. As long ago as 1944, Lindemann published an observation of the psychological aftermath of the Coconut Grove nightclub fire in Boston. Besides the obvious involvement psychologists have in attempting to relieve distress of victims, disasters have a relationship to several important psychological constructs. Disasters allow psychologists to perceive the operation of trauma on emotional functioning, an operation which mental health practitioners as far back as Freud have been interested in understanding. Stress research is a central and crucial explanatory factor in many fields of psychology, especially community psychology, which considers stress the central ingredient to the formation of psychopathology (e.g., Albee, 1997; Dohrenwend, 1998). There is an ethical limit to the extent that stress can be manipulated in the laboratory, and disasters allow psychologists the opportunity to observe how extreme stress impacts individuals and groups.
Because of psychology's interest in trauma and stress, its definition of disaster has differed somewhat from that employed in other fields. In the 1970's, after the Vietnam War and the discovery of its impact on veterans, and after the discovery of the long-term effects of child sexual abuse, the mental health field conceptualized a disorder specifically related to the consequences of trauma, Post Traumatic Stress Disorder (PTSD) (American Psychiatric Association, 2000). We will define PTSD later in the chapter. Here we are making the point that because of the interest by psychologists in PTSD, there has been some blurring between the concepts of victimization from any source and victimization from disaster. For instance, vulnerability factors to PTSD in victims of an earthquake may be similar to vulnerability factors to PTSD in victims of rape, and effective treatments may also be similar, so that studying a broader group of victims may be useful in understanding disasters.
In spite of many differences in opinion (e.g., Quarantelli, 1998), the definition of disaster in use in this chapter agrees with that of most psychologists (e.g., Barton, 1969; Norris, Friedman, Watson, Byrne, Diaz & Kaniaty, 2002) who regard disasters as involving an unexpected or uncontrollable event rather than a long-term experience. That is, a disaster is something that could happen within a war (e.g., My Lai, or many other less well-known examples) rather than the war itself, or Three Mile Island rather than Love Canal. These examples illustrate the difficulty with the distinction, and some researchers think that our concept of disaster should include chronic disaster (Couch & Kroll-Smith, 1985). Dynes (2004) has argued that social scientists need to expand their definition of disaster to encompass events like war, genocide, and refugee experiences that are critical in third world countries.
Disasters are also usually viewed as a collective experience, excluding personal disasters like sexual abuse or automobile accidents, unless these involve a large number of people. Again, the dividing line can be unclear. The type of event, with its various dimensions, can affect our perceptions. We might not consider an automobile accident that killed 13 people to be a disaster, even if many others were involved or witnessed it, but the killing of 13 in the shootings at Columbine certainly qualifies.
With the passage of time, study of disasters has become less descriptive and more quantitative, attempting to resolve some of the methodological problems of this research. The focus has moved from the question of whether there are significant long-term psychological impacts of disasters, to studying the types of impact that occur and what factors in the disaster and in the individual increase the likelihood of emotional damage. Interventions to assist victims have been developed. Most recently, there has been more focus on the effectiveness of these interventions. This chapter will explore in turn each of these areas: methodology of disaster research; extent of psychological impact of disasters; types of psychological sequelae; damaging aspects of disaster; vulnerability factors; psychological interventions for victims; and the effectiveness of these interventions.
Methodology of Disaster Research
Early studies of disaster tended to be descriptive. Lifton (1967) described the emotional impacts of Hiroshima, Coles (1967) portrayed the effect of political disaster on children (1967), and Erikson (1976) painted the picture of the aftermath of the Buffalo Creek floods in West Virginia. While some researchers (e.g., Edelstein, 2004) still favor a qualitative approach, most psychological disaster research today tends to be quantitative.
Problems exist for the social scientist who wishes to study disaster. Experimental design requires random assignment of participants to experimental and control conditions. Even if a mad scientist wanted to conduct such an experiment, controlling a disaster is an oxymoron. Disaster research can only attain the status of quasi-experimental design, with comparison groups, not controls. Since disasters occur unpredictably, pre-test data on victims are usually not available. Psychologists called into a disaster are usually there to provide help. Researchers can seldom obtain access to the disaster at its onset, and if they do find access, the exigencies of the situation usually preclude administration of standard instruments in a standardized fashion. Victims usually have no motive to participate in research, and follow-up studies are often difficult to arrange. Samples of victims vary from those directly impacted, to rescue workers, to the families of the bereaved. It is difficult to compare Western victims to those from third world countries, as their circumstances and resources are so different, and for the same reasons it is difficult to compare victims from different ethnic groups within a culture.
Disasters also vary widely in the amount and the nature of the stress they involve: duration; loss of life; personal injury, or injury to loved ones; property damage; terror; helplessness; gruesome sights, sounds and smells; dislocation from one's home; availability of social support – all these factors may differ in a flood as contrasted to an earthquake, or between one flood and another, or between one victim's and another's experience of the same flood. One special differentiation between types of disaster is the natural vs. the technological, or human-caused, disaster. Natural disasters tend to involve lack of control over natural forces, like wind, that we expect to be uncontrollable, while technological disasters can be less defined, especially if they include toxic exposure, and can involve a loss of control over an area of life in which we expect control, like drinking water (Baum, Fleming, & Davidson, 1983). Terrorism is a special form of technological disaster, and the most recent addition to the typology of disaster (Ursano, Fullerton, & Norwood, 2003).
As psychologists conducted more disaster research, they began to develop standardized measures, beginning with the Impact of Events Scale (Horowitz, Wilner & Alvarez, 1979). Many measures have been devised to diagnose post-traumatic stress disorder (PTSD), one of a number of psychological consequences of disaster. The National Center for PTSD (2003) currently lists 15 adult PTSD self-report measures, 4 interview measures, and 9 measures for children. Obviously, many other standardized measures of other types of psychopathology have been administered. Measures have also been developed to identify vulnerability factors and intervening variables, e.g., the Peritraumatic Dissociation Experiences Questionnaire (Marmar, Weiss, Schlenger et al., 1994), and World Assumptions Scale (Janoff-Bulman, 1985).
Comparison groups, if not actual control groups, were introduced early into the research. A step forward in the confusing array of studies on different disasters with different samples and different methods came with the meta-analysis of Rubonis and Bickman (1991), which found small but consistent post-disaster effects on levels of psychopathology across different types of study and types of disaster.
Robins, Fischbach, Smith, Cottler, Solomon, and Goldring (1986) seized upon a fortuitous (or infortuitous) series of events at the Times Beach site in Missouri. Interviews had taken place in that area for the Epidemiological Catchment Area study which documented the prevalence of psychiatric problems in the country. Then the area was struck both by floods and the discovery of dioxin. This allowed the comparison of the effects of a natural and a technological disaster, with pre-test information available for a sample of the victims, and documented that change had occurred.
When Norris et al. reviewed the disaster literature in 2002, she found six other studies that were able to obtain true pre-test measures for their samples. Comparison of types of disaster exposure for the same sample remains rare. Norris et al. also report that many recent disaster studies used follow-up formats and probability sampling methods. In short, methodology has improved dramatically, and conclusions can comfortably be drawn about the psychological impact of disasters.
Extent of Psychological Impact of Disasters
As noted, Rubonis and Bickman found in their 1991 meta-analysis consistent but small post-disaster effects upon psychopathology. Many other review chapters and articles have been written (e.g., Gibbs, 1989, 1991, Green & Solomon, 1995; Katz, Pellegrino, Pandya, Ng, & DeList, 2002; Sundin & Horowitz, 2003), concluding that post-disaster effects are greater and more pervasive than Rubonis and Bickman's inferences. The most thorough recent review is that of Norris et al. (2002a). The authors analyzed 160 different disaster studies, with a total of over 60,000 participants, and did not conduct a meta-analysis because of the difficulty in deriving effect sizes from descriptive studies. Using a rough four point scale to rate level of pathology, they found that only about 10 of studies found minimal impairment, about half the studies found moderate impairment, and the remaining 40% found severe or very severe impairment. Severe impairment was equivalent to rates of psychopathology in the participants of between 25 and 50%.
For many years there has been a debate over whether the effect of disaster on mental health was important. One side of the debate came from the sociological point of view (e.g., Quarantelli & Dynes, 1985), which focused on the adaptive nature of community response, both in the immediate aftermath of a disaster and in most people's long-term response. The majority of people function adaptively during and after a disaster, and the old notion (Kinston & Rosser, 1974) that individuals will experience panic, wander aimlessly and be dependent has been shown to be untrue (Wenger, Dykes, Sebok, & Neff, 1975). But that is a different matter from focusing on the toll that the disaster takes on some individuals. There is so much evidence now of the damage to individuals, that to our minds, the debate has been resolved. Norris, Friedman, & Watson (2002b) conclude that the field does not need new studies indicating that disaster causes serious psychopathology; we know this to be the case. Instead, we should focus on understanding what aspects of disaster are most devastating, and what characteristics of individuals make them vulnerable, issues we will address later in the chapter.
Forms of psychopathology resulting from disasters.
If it is clear that disasters cause psychopathology, it is less clear what form that psychopathology takes. Since the mental health profession developed the PTSD diagnosis, PTSD has been the main focus of research on the aftermath of disaster. The criteria for PTSD include (APA, 2000): 1) having been exposed to a traumatic and fearful event; 2) re-experiencing the traumatic event, usually in flashbacks or nightmares; 3) avoidance of situations and stimuli that could reawaken the trauma, for example, numbing one's feelings or withdrawing from others; and 4) increased level of arousal, for instance, sleep difficulties, irritability, and concentration problems.
Norris et al. (2002a) reported that 68% of their research samples assessed for and found PTSD in disaster victims. The second most common psychiatric problem was depression, found in 36% of the samples. Anxiety in various forms was shown in 32% of the samples. Health concerns were also often present (23% of the samples). It was not usually clear whether victims' health concerns were realistic, or were based on somaticizing the stress of the experience (North 2002). Alcoholism and drug abuse were not often investigated but when they were, levels of abuse have been found to rise after disasters.
What is not clear from the above figures is what the actual rate of various psychopathologies might be if each study had assessed for all of them. Norris et al. (2002b) recommend that all disaster researchers use a standard measure of psychopathology so that it can be more clearly determined which disorders are linked to undergoing disaster.
Victimization, primarily child physical and sexual abuse, has been shown to lead to other diagnoses beyond the ones investigated in disasters. These include schizophrenia and other psychoses (Neria, Bromet, Sievers, Lavelle, & Fochtmann, 2002), dissociative disorders (Coons & Milstein, 1986) and borderline personality disorder (Herman, Perry & Van der Kolk, 1989). None of these diagnoses has been investigated to see if higher rates result after disaster, although dissociative symptoms have been reported during and after some disasters (Marmar, C. R., Weiss, D. S., Metzler, & DeLucchi, 1996; Weiss, Marmar, Metzler, & Ronfeldt, 1996) and can be part of the avoidance criterion of PTSD (APA, 2000). It would be valuable to look at long-term vulnerabilities of childhood victims of disaster to these disorders. Little research of any kind has been conducted looking at long-term consequences of disasters for children.
An issue that has been discussed in the literature is whether symptoms of other disorders found after disasters are part of the PTSD syndrome or whether they are independent consequences. There are several possible explanations for the overlap that often is observed. Symptoms within diagnoses do overlap, symptoms of other diagnoses could be sub-clinical cases of PTSD, PTSD could increase vulnerability to other diagnoses, and other diagnoses could increase vulnerability to PTSD (McMillen, North, Mosley and Smith, 2002). In particular, the fact that depression and PTSD are both common consequences of disaster is of interest. Greening, Stoppelbein, & Docter (2002) conducted an interesting study in which they looked at attributions for the negative outcomes of the Northridge earthquake. Victims who made what have been labelled depressogenic attributions, seeing negative outcomes as related to internal, stable and global causes (Abramson, Seligman, & Teasdale, 1978), were more likely to develop depressive symptoms, but not PTSD symptoms. Livanou, Basoglu, Salcioglu and Calender (2002) looked at PTSD and depression as outcomes of the Turkish 1999 earthquake, and found that there were different predictors for each. Research into the relationship between different outcomes of disaster is continuing, but the lack of solid findings points out that we know little about the actual mechanism of how symptoms are caused by disaster stress.
Aspects of disaster which contribute to psychopathology
In general, the nature of the disaster and the extent of the trauma it wreaks are more predictive of the extent of psychopathology that follows than are characteristics of the victims (Sundin & Horowitz, 2003). The more stressful the disaster experience, it appears, the more negative the consequences, but it is not always possible to identify which of the many factors within a disaster make it more stressful. Theorists have identified the following as important characteristics: mass violence (Norris et al., 2002); the experience of terror and horror (Bolin, 1985); duration of the disaster (Baum & Davidson, 1985, Bolin, 1985); and the amount of unpredictability and lack of control (Baum & Davidson, 1985; Thoits, 1983).
First responders and disaster workers are at special risk for PTSD and other negative emotional consequences of disaster (Gibbs, Lachenmeyer, Broska, and Deucher, 1996; Norris, 2002a) . This vulnerability has usually been perceived to be related to the experience of the work rather than to any inherent vulnerability factors, as often people choosing these professions have high levels of emotional hardiness. Looking at disaster workers who dealt with the aftermath of the World Trade Center disaster of September 11th provides an example. Working with dead bodies and body parts after a major disaster is something that almost everyone finds extraordinarily stressful, and perhaps the experience could be said to define horror. Disaster workers' experience of the disaster is often more long-term than that of other victims, as for instance the long term digging out after September 11th. In addition, the experience of helplessness and lack of control is prevalent, as workers searched for but were unable to find identifiable bodies.
Psychologists have many theories about what causes the disorders of PTSD, depression, anxiety reactions, etc. (e.g., Barlow, 2000), but little conclusive about what it is exactly about a disaster that leads to emotional damage.
As we have mentioned, most psychologists identify stress as a leading cause of psychopathology, but theories as to how stress affects its victims are varied. Some focus on the physiological overload of stress (e.g., Selye, 1976), some on the unpredictability and uncontrollability of stress (e.g., Kelly, 1955) and some on the conditioning that takes place between a frightening stressor and other aspects of life, with a resulting avoidance of stimuli that are reminders (Mowrer, 1960). Losses in a disaster, of other people, of material goods, of one's own health and security, are also critical (Nolen-Hoeksema, 1990). Some theorists focus on the shift in cognitions that take place after a disaster. Janoff-Bulman and Frieze (1983) speculated that cognitions shift after a disaster. The individual asks "Why me?" and the answer involves a change in one's sense of invulnerability, in the world's predictability, and in one's own worth.
Research has identified a number of characteristics of victims that make them more vulnerable to disaster effects. Vulnerability factors include, but are not limited to, socioeconomic status (SES), available resources, previous level of psychopathology, age, social/family factors, gender and ethnicity.
Regarding SES, Norris et al. (2002a) found that thirteen of fourteen samples which investigated socioeconomic status and disaster outcome found lower socioeconomic status to be associated with increased post-disaster distress. Studies included a wide range of disasters: an air disaster (Epstein, Fullerton, & Ursano, 1998), an industrial disaster (Vila, Witowski, & Tondini, 2001), floods (Ginexi, Weihs, Simmens, & Hoyt, 2000), and an earthquake (Lewin, Carr, & Webster, 1998). Individuals who live in poverty tend to have fewer resources available to them to attenuate the effects of disaster.
Pre-existing psychopathology is a risk factor for developing psychopathology related to a trauma (Norris et al., 2002a) in that individuals who suffer from a psychological disorder are more susceptible to further distress in the aftermath of a disaster. For example, pre-disaster anxiety disorders (Asarnow et al., 1999), depression (Knight, Gatz & Heller, 2000), and suicidal ideation (Warheit, Zimmerman & Khoury, 1996) were found to increase the likelihood of post-disaster psychopathology.
In terms of age, Norris et al. (2002a) noted that middle-aged adults appear to be the group most affected by disasters. This age group may have more burdens and stresses (Thompson, Norris & Hanacek, 1993), such as caring and providing support for a family, that may be amplified in the aftermath of a disaster.
Social network characteristics influence vulnerability. For example, a lack of perceived (Bromet, 1982; Dougall, Hyman & Hayward, 2001) or received (Sanchez, Korbin & Viscarra, 1995; Udwin, Boyle & Yule, 2000) social support may lead to greater post-disaster distress.
These risk factors do not operate in isolation. Any single factor is often interrelated with others. We will illustrate this complex interaction within the context of two variables: gender and minority or third world ethnicity, both of which Norris et al. (2002a) in their review cite as among the most robust of vulnerability factors.
Gender. Norris et al. (2002a) stated that in 94% of 49 studies which investigated the issue, female survivors of disaster were more seriously affected than were males. There are several possible explanations for this difference. For example, as mentioned in the previous paragraph, low socioeconomic status is a risk factor for post-disaster psychopathology, and women more often live in poverty than men (Belle, 2000).
The gender difference may be in part explained by differences we often observe between the genders in the way psychological distress is expressed. In general, women are more likely than men to acknowledge psychological symptoms and to report them (Nolen-Hoeksema, 1990). After a disaster, males may suppress feelings of psychological distress because of the expectation that men must be strong and capable (Wolfe & Kimerling, 1997). As discussed in a previous section, the most commonly investigated post-disaster reactions are PTSD, depression, and other forms of anxiety. Substance abuse and other acting out behaviors, such as interpersonal violence, are seldom assessed. Men are more likely to express psychological distress through these kinds of behaviors, rather than reporting neurotic-type symptoms like depression and anxiety (Myers, Weissman, Tischler, et al., 1984).
Women have higher pre-disaster rates of depression and most anxiety disorders than men (Myers et al., 1984), putting them at risk for disaster-related distress. Furthermore, there may be some experiences that women are more likely to have that may contribute to the development of PTSD post-disaster. The experience of rape and sexual assault is higher among women than men (Kessler, Sonnega, Bromet, Hughes & Nelson, 1995), and it has been shown that when compared with other forms of trauma, unwanted sexual contact is more likely to result in PTSD ( Breslau, Davis & Andreski, 1997; Kessler et al., 1995). Pulcino et al. (2003) found that the experience of previous unwanted sexual contact increased a woman’s likelihood of endorsing PTSD symptoms after the September 11th attacks by 33%.