EVELYN J. BROMET

The psychological aftermath of three radiation catastrophes is reviewed, namely, the bombings of Hiroshima and Nagasaki, the accident at the Three Mile Island nuclear power plant, and the accident at the Chernobyl nuclear power plant. The major psychiatric and psychological symptoms found in the affected populations were a preoccupation with somatic symptoms, clinical depression, anxiety/fear of future health effects, neurasthenia (fatigue, weakness, dizziness, headaches), and post-traumatic stress symptoms. The symptoms were found to persist for many years after these events. Risk factors included exposure to death and illness, physical and social stigma, community disruption, and personal vulnerability factors such as gender and a history of psychopathology. The need for systematic, psychiatric epidemiologic research among Hiroshima/Nagasaki survivors and long-term follow-up studies of all three cataclysms is emphasized.

From a public health perspective, the psychological effects of nuclear catastrophes may be equally, if not more, prevalent than their physical health consequences. The bombing of Hiroshima and Nagasaki was an extraordinary trauma for the survivors. Yet, epidemiologic research into its mental health consequences is virtually nonexistent. By contrast, during the last few decades, there have been a growing number of epidemiologic studies of the psychiatric, psychological, and social effects of natural and technological disasters and war-related traumas (Bromet and

Schulberg, 1987; Solomon, 1989; Kulka et al., 1990; Bromet and Dew, 1995). A recent meta-analysis of this body of disaster research concluded that the average excess psychological morbidity was 17% (Rubonis and Bickman, 1991). However, some of the most horrific events in modern history, such as the bombing of Hiroshima and Nagasaki and other disasters occurring in third world and eastern European countries, had not been studied and hence were not included in the review. It seems reasonable to surmise that the average excess morbidity rate of 17% might have been higher had such events been considered.

Psychiatric epidemiology is concerned with understanding the distribution of psychiatric disorders in the population and the risk factors associated with their onset and course (Bromet and Parkinson, 1992). Mental health disaster research is one important component of this field. The impetus and basic tools needed to advance research in psychiatric epidemiology stemmed from several significant experiences that occurred during World War II. First, psychologists in the armed forces developed a psychological symptom questionnaire, the Neuropsychiatric Screening Adjunct, to screen recruits. More recruits were rejected because of psychological impairment than because of physical impairment. It was therefore obvious that treated cases represented the tip of the iceberg. Thus, following World War II, a number of prevalence studies were carried out to determine the true prevalence of mental disorder in the population, starting with symptom-based research characterizing the 1950s to 1970s (Dohrenwend and Dohrenwend, 1982) and culminating in the recent disorder-based research beginning around 1980 (Robins and Regier, 1991; Kessler et al., 1994; Tsuang et al., 1995). Another event that shaped the focus of much psychiatric epidemiologic work was the observation that many soldiers suffered from "shell shock," or post-traumatic stress syndromes, from the extraordinary stresses to which they were exposed—even though they had screened "negative" for psychological disorder upon recruitment. Disentangling the links between stress and mental disorder became an important area of research and was for the most part the conceptual impetus for the large-scale post-war studies, such as the Midtown Manhattan and the Stirling County studies (Leighton et al., 1963; Robins and Regier, 1991). Unfortunately, the expansion of research in psychiatric epidemiology, including disaster research, did not extend to the group of people who experienced one of the most horrific stressors in history, the survivors of Hiroshima and Nagasaki.

Radiation catastrophes share an important underlying dimension, namely, the perception of risk to health (Slovic, 1991), a perception that for many survivors can turn into an unresolvable fear. This fear may be reinforced when cancers and other illnesses occur among survivors and are attributed, rightly or wrongly, to the radiation exposure. This chapter reviews findings from three unique radiation catastrophes, Hiroshima, Three Mile Island, and Chernobyl. Each differed in form, historical period, sociocultural context, and immediate consequences. Each also differed in the research perspectives and methods that were brought to bear

to understand their psychologic effects. Yet the parallels in the findings on psychological reactions to these events are striking. The findings from these three events clearly demonstrate that the psychological legacy of collective radiation exposure—unremitting fear of radiation-related illness—is independent of the extent of exposure; or to put this in statistical language, the psychological effects are poorly correlated with the exposure dose. The following section presents a review of the existing studies of these events, as summarized in Table 18.1. Next, three issues are explored: an integration of the findings on the types of mental health problems arising after radiation exposure; the risk factors associated with negative psychological effects; and the identification of factors that could modify the impact of this exposure.

Although there are some Japanese-language publications addressing the psychological sequelae of the bombings of Hiroshima and Nagasaki in 1945, only three English-language publications were found, and two were quasi-epidemiological. The first, by Misao and colleagues (1961), described the subjective complaints of 356 survivors (177 females, 179 males), as well as symptoms reported on the Cornell Medical Index completed by 345 survivors (149 female, 196 male). The actual source of the samples was not described in the paper. The findings from the Cornell Medical Index were compared with data from a normal sample of clerks and medical outpatients, but no specific information about the origins of these samples and their data was given. Overall, the authors suggested that the rates of subjective complaints (>50%) and anxiety about possible "A-bomb" diseases (>75%) were elevated. The Cornell Medical Index scores were also noted to be higher in survivors than in controls, although specific data on the controls were not presented.

The second quasi-epidemiologic paper presented findings on the suicide rate from 1950 to 1966 of members of the Life Span Study (n=109,000 exposed and non-exposed individuals) and on symptom responses to a questionnaire modeled after the Cornell Medical Index, which was administered to 10,522 members of the Adult Health Study (AHS) from 1962 to 1965 (Yamada et al., 1991). While no significant difference in the suicide rate was found, symptoms of fatigue, anxiety, and depression were significantly higher in the exposed compared to non-exposed group.

The third study was the detailed report by Lifton (1967) based on in-depth, guided interviews with survivors that he personally conducted in 1962. The generalizability of the findings is open to question because of both the unstructured format of the interview and the nature of the sample, that is, survivors from an unspecified population pool willing to talk at length about highly personal matters with a psychoanalytically oriented American psychiatrist. Nevertheless, this work

provided a rich resource regarding the psychological experiences of survivors and helped shape the development of current psychiatric nosology for post-traumatic stress disorder. In particular, Lifton's descriptions of psychic numbing (originally referred to as "psychological closure")—that is, the loss of feeling from the extreme agony that resulted from witnessing mass death and dying and being unable to respond to calls for help—provided important insights into this poorly understood phenomenon. This phenomenon, as well as other cardinal symptoms Lifton described, ultimately became criterion symptoms for the classification of post-traumatic stress disorder in the American Psychiatric Association's Diagnostic and Statistical Manual starting with the 1980 edition. The concepts of intrusive and avoidant thinking also were operationalized in the Impact of Events Scale (Horowitz, 1990), which has been widely used in disaster research and studies of war-related traumas.

Although the accident at the TMI nuclear power plant in central Pennsylvania was regarded at the time as the "worst … in the history of commercial nuclear power generation" (Kemeny, 1979, p. 35), radiation health experts concluded that the emission was too small to increase the risk of cancer in the population living near the plant (Upton, 1981; Upton et al., 1992). Recent epidemiologic findings on cancer seem to support that conclusion (Hatch et al., 1990). At the same time, these experts asserted that there was a significant health impact on "anxiety and stress." The Task Force on Behavioral Effects of the President's Commission on the Accident at Three Mile Island concluded, however, that the mental health effects were "short-lived,'' diminishing sharply by the fall of 1979 (Dohrenwend et al., 1979; Kemeny, 1979; Dohrenwend et al., 1981). This conclusion was subsequently contradicted by several longitudinal studies, including a three-year follow-up of 254 residents who had participated in research subsumed within the Task Force report (Goldsteen et al., 1989). Similarly, high levels of distress were also reported by Houts et al. (1988) among residents participating in telephone interviews by the Pennsylvania Department of Health in January 1980, and by Baum et al. (1993) in a six-year, longitudinal study of a panel sample from the five-mile radius of TMI. The comparison groups in the latter study included residents living near another nuclear reactor, a coal-fired plant, and more than five miles from any power plant. The Baum et al. study found higher rates of a very wide range of symptom areas up to six years after the accident, including somatic complaints, anxiety, and depression; PTSD symptoms of hyperarousal, frequent and bothersome intrusive thoughts about the accident, and avoidance of reminders of it; and physiological symptoms including elevations in blood pressure and in levels of urinary norepinephrine, epinephrine, and cortisol.

Our findings also contradicted the early conclusions of the Task Force report (e.g., Bromet, 1989; Bromet et al., 1990; Bromet, 1991; Dew and Bromet, 1993). This study focused on clinical and subclinical depression and anxiety syndromes in mothers of young children and psychiatric outpatients (studied for one year) who lived within 10 miles of TMI, as well as workers employed there when the accident occurred (Bromet, 1989). Comparison groups were sampled from around a similar nuclear power plant and a coal-fired generating plant. The group who proved most vulnerable to the stresses associated with the accident were the mothers of young children. In four waves of face-to-face interviews conducted from 1979 to 1982, we found that (1) their psychological symptoms were persistently elevated; (2) perceptions of danger became a stronger predictor of distress as time elapsed; (3) pre-TMI accident depression and/or anxiety disorder was the most important prognostic indicator; (4) social support did not moderate the mental health effects of the TMI accident; and (5) only half of the mothers with clinical depression sought help for their episodes (e.g., Bromet et al., 1982; Dew et al., 1987a; Dew et al., 1988). We also found no evidence that the mental health of the mothers' preschool (Cornely and Bromet, 1986) or school-aged children (Bromet et al., 1984) differed

significantly from that of comparison site children, but a positive family milieu was found to buffer the stress-mental health relationships in the TMI school-aged sample. After collecting two additional mail-back questionnaires from the TMI mothers in 1985 and 1989, we further observed that the restart of the undamaged reactor had an adverse psychological effect (Dew et al., 1987b) and that by 1989, approximately 35% of the mothers had consistently high levels of distress over the ten-year period of observation (Dew and Bromet, 1993). The distress was associated with continued concerns about TMI and with living within five miles of the plant when the accident occurred (Bromet et al., 1990). In fact, the evacuation experience and perception that TMI was dangerous in 1979 "remained among the strongest contributors to long-term distress, even after other important background characteristics were statistically controlled" (p. 54).

Thus, while TMI proved to be among the least dangerous of the nuclear catastrophes since World War II, it is perhaps the best studied and hence best understood.

Like Hiroshima and Nagasaki, the effects of the explosion at Chernobyl are complicated to evaluate. Thousands of families living near the Chernobyl plant were exposed to radiation. The evacuation of the 30-kilometer zone was chaotic and at times brutal, and the settings to which evacuees were relocated were openly unreceptive and even hostile. Anecdotes about persistent fears of radiation effects (given the label of "radiophobia" by Russian professionals) and chronic stress associated with feelings of stigma (similar to Lifton's descriptions of social stigma) have frequently been reported. While there have been presentations at international conferences (see Ginzburg and Reis, 1991 for a synopsis of one such meeting) and reports in Russian-language research journals, the only published English-language studies to date of the psychological well-being of community residents exposed to radiation from Chernobyl were a small section of the International Chernobyl Project technical report from the International Atomic Energy Agency (International Atomic Energy Agency, 1991), a study of a small exposed village (Viinamaki et al., 1995), and a general population study in Gomel, Belarus (Havenaar et al., 1996). The first study was the largest (perhaps only) epidemiologic study to date that evaluated the health status of five groups who remained in rural contaminated communities (2-year-olds born in 1988; 5-year-olds born in 1986; 10-year-olds born in 1980; 40-year-olds born in 1950; and 60-year-olds born in 1930). The team concluded that these populations were not significantly different from controls on hematological, thyroid, and general health problems (except abdominal problems), but that 3–5% of all children had health problems that required treatment. However, the study excluded evacuees and cleanup workers ("liquidators") whose exposure after the accident was much greater. Although psychological stress was reported to be more prevalent in the settlements contaminated by radiation, the study did not include standard psychological scales to measure distress. The second and third studies found more impairment (based on responses to the 12-item

General Health Questionnaire) in exposed groups than in controls (Viinamaki et al., 1995; Havenaar et al., 1996).

There are three ongoing, western-based studies of populations exposed to Chernobyl which will ultimately yield important new data on the psychological consequences of this event. The first is the neurological and neuropsychological screening of exposed children who were in utero at the time of the accident (World Health Organization, 1995). A preliminary report in Russian appears to suggest that compared to controls, the mothers of these children have higher psychological (stress) symptom scores, and the children show poorer performance on some of the neuropsychological measures. The second ongoing study is a clinical assessment of liquidators (cleanup workers) living near Moscow (Lasko et al., 1993). The third study addresses the mental health effects of the stress of the Chernobyl experience on evacuee children living in Kiev (Bromet et al., 1994). The latter study was funded in September 1995 by the National Institute of Mental Health.

Two other reports about the psychological sequelae of Chernobyl should be noted. One was based on a study of a small sample of exposed adolescent girls, subsequently brought to Israel, who showed high levels of somatization but no detectable underlying medical conditions at the time of evaluation (personal communication, Lerner and Zilber, 1992). The second study was conducted by two Ukrainian investigators in spring 1992 and focused on the self-reported health status of 249 9^th to 11^th grade students evacuated to Kiev, along with several comparison groups (personal communication, Panina and Golovakha, 1992). Compared to controls drawn from Kiev, evacuated adolescents were more likely to report frequent headaches, chronic illness, poorer overall health, and lower satisfaction with personal health. On the other hand, they did not report a greater numbers of the school day, or higher anxiety symptoms [assessed with the Spielberger (1973) scale]. Compared to controls living in Slavuta, a town built to house workers from the Khmelnitsk nuclear power station, the evacuees manifested significantly greater anxiety but were otherwise similar in their psychological profiles.

The most common symptoms reported across the studies described above were somatization, depression, post-traumatic stress symptoms, anger/hostility, anxiety (particularly about future health), and, similar to somatization, a combination of symptoms that was once labeled neurasthenia—namely, fatigue, irritability, weakness, and headaches. New and unfamiliar terms have been used to describe the psychological and physical symptoms of radiation survivors, including A-bomb neurosis in Japan and vascular dystony (Stiehm, 1992) and radiophobia (Ginzburg and Reis, 1991) in the former USSR. Table 18.2 summarizes the symptoms and conditions reported in the described studies.

While the specific rates of disorder vary because of methodological differences in study design, measures, samples, and timing of data collection, the most remarkable aspect about the findings is the intractable nature, or persistence, of these symptoms. In our TMI research, as noted above, 35% of mothers of young children continued to show elevated rates of distress over a ten-year period (Dew and Bromet, 1993). Similarly, the temporal persistence of fearful perceptions was highly significant (Dew et al., 1987c). The Misao et al. (1961) study of A-bomb survivors found that more than 50% reported fatigue, headache, anxiety, and somatic

complaints 14 years later. When Yamada et al. (1991) assessed AHS members 17 years later, they found elevated rates of anxiety, fear of disease, fatigue, and depression. Finally, the IAEA finding of elevated rates of distress was based on data collected five years after the catastrophe took place (International Atomic Energy Agency, 1991).

Thus a strong case can be made for extended, long-term follow-up research focused on the psychological effects of these catastrophic events. As Newsweek magazine (July 27, 1995) stated about the survivors of Hiroshima and Nagasaki, "Nothing will ever remove the physical and psychic pain of the hibakusha [A-bomb survivors]."

A number of personal and environmental risk factors have been found to elevate the rate of psychological disorder from both technological and natural disasters (Table 18.3). These include greater involvement and/or injury, exposure to death and dying—particularly in family members, exposure to cancer, physical and social stigma, community disruption, economic decline, unresolved fear of the disaster's future effects, and personal vulnerability factors, particularly female gender and having a personal history of psychiatric disorder (Baum, 1987; Bromet and Schulberg, 1987; Solomon, 1992; Dew and Bromet, 1993). It can be hypothesized that the less severe the disaster, the stronger will be the contribution of personal vulnerability factors (Davidson and Foa, 1993). Thus, at TMI, our research showed that personal psychiatric history was the strongest predictor of post-TMI psychopathology. Conceivably, disaster-related stressors may override the moderating impact of these personal vulnerability factors in the A-bomb survivors. Although there are few studies of the specificity of these risk factors for different outcomes (Solomon et al., 1993), it would be important to understand which risk factors predict different types of outcomes so that more specific prevention efforts can be designed.

A number of variables have been suggested that can be conceptualized as modifying factors. That is, their presence may either influence the magnitude of or moderate the effects of the stress from these events. One such factor is media coverage. As early as 1961, Misao et al. claimed that A-bomb survivors learned about symptom effects from the mass media. Specifically, they asserted that the media's lack of caution in describing the sequelae was in part responsible for the level of anxiety in the survivors. At a 1990 symposium on psychological effects of radiation accidents, the President of Radiation Management Consultants introduced a session by urging scientists to "determine how, and how much, stress is related to the media's misrepresentation of the accident and the long-term effects of radiation" (Linnemann, 1991, p. 58). While it is easy to attribute the appearance

or persistence of symptoms to miscommunication on the part of the mass media, it should be recognized that scientists often made little effort to communicate their findings directly to the affected populations. If we assumed this responsibility more assiduously, the impact of inaccurate communications from the mass media might be lessened.

Social psychologists and psychiatric epidemiologists have also enumerated a number of other variables that could attenuate the effect of the stress. These include cognitive control over intrusive thoughts which can be gained from professional treatment (Baum, 1990; Baum et al., 1993), practical support from government agencies (in the form of medical benefits, social security, and housing; Solomon et al., 1987), and social support from family, social network members, health care and religious professionals (Fleming et al., 1982; Baum and Fleming, 1993). The issue of social support is particularly important. After the A-bomb, the hibakusha tended to marry each other and to maintain lower socioeconomic status because they were perceived as undesirable for marriage, child-bearing, and employment. Similarly, evacuees from the 30-kilometer zone around Chernobyl were stigmatized and received with hostility by the communities where they were relocated because they were feared, and looked upon as sickly. To some extent, the persistence of psychological symptoms might be attributed to these secondary adversities occurring in the aftermath of these catastrophes.

Better data are needed about the prevalence and risk factors for different types of psychiatric disturbance following the radiation catastrophes of the twentieth century. While it appears that certain types of symptoms may be widespread, their

true prevalence is unknown. Because these events have been shown to have long-lasting effects, it is still timely to initiate a long-term outcome study of A-bomb survivors as well as individuals affected by Chernobyl and TMI. Data on risk factors, as well as factors that protect against deleterious mental health outcomes, would help us design targeted and useful intervention strategies. As nuclear power plants continue to age and plans for radioactive waste disposal and nuclear weapons disposal are enacted, these populations become increasingly important to evaluate. The long-term threat to health is a reality that must be understood. By comparing the long-term effects of several radiation-related events, we may be able to elucidate the common effects of exposure, or threatened exposure, to ionizing radiation (Collins and de Carvalho, 1993). Thus, it is crucial that research with both practical and theoretical underpinnings be designed to address the natural history of psychological response to this form of extreme stress. These studies need to address both the survivors and their offspring, because the fear extends beyond the survivor and into the next generation.

I wish to thank Julia Bromet Pilkonis for her exhaustive search and thoughtful summary of the literature on the effects of the bombings of Hiroshima and Nagasaki. Her assistance in this area was invaluable.