MedicineLakex

 

Acp_566 194.200

Acta Psychiatr Scand 2005: 112: 194–200 Copyright
2005 Blackwell Munksgaard All rights reserved ACTA PSYCHIATRICA Psychological aftermath of the Lviv air showdisaster: a prospective controlled study Bromet EJ, Havenaar JM, Gluzman SF, Tintle NL. Psychological E. J. Bromet1, J. M. Havenaar2, aftermath of the Lviv air show disaster: a prospective controlled study.
S. F. Gluzman3, N. L. Tintle1 Acta Psychiatr Scand 2005: 112: 194–200.
2005 Blackwell Munksgaard.
1Department of Psychiatry, State University of New Yorkat Stony Brook, NY, USA, 2Department of Psychiatry, Objective: To investigate the psychological aftermath of an air show Vrije Universiteit Amsterdam, Amsterdam, the disaster using prospectively obtained epidemiologic data.
Netherlands and 3Ukrainian Psychiatric Association, Method: Participants in a recently completed epidemiologic mental health survey in Lviv (disaster site) and controls from western Ukrainewere interviewed shortly before and 6 months after a gruesome airshow disaster.
Results: The Lviv group reported more psychopathology and post-traumatic stress symptom severity, but less anomie than controls.
Somatization symptoms were similar in the two groups. Predisastermental health and postdisaster threat were the strongest risk factorswhile demographic characteristics, emotional support, and repeated Key words: disaster; post-traumatic stress; television viewing of the event were only weakly associated with somatization; anomie; prospective postdisaster mental health.
Evelyn J. Bromet, Department of Psychiatry, Putnam Conclusion: This is the first prospective study to find a significantly Hall-South Campus, Stony Brook, NY 11794-8790, USA.
higher rate of disorder as well as post-traumatic stress disorder symptomatology after a disaster. The risk factor findings suggestavenues for targeting postdisaster interventions.
Accepted for publication April 11, 2005 with a history of psychopathology, and most importantly, the most severely exposed (1, 2, 11).
Research on the psychological consequences of Although disasters are conceptualized as Ônatural disasters has grown exponentially in the last two experiments,Õ only a handful of studies have used a decades (1). Overall, these studies place the 1-year pre–post design and hence contain unbiased pre- disaster-attributable prevalence of psychological disaster morbidity data. The absence of such morbidity at about 20% (2), but the range of predisaster baseline data places constraints on the probable cases may exceed 50% (e.g. 3). In inferences that can be drawn about postdisaster addition to psychiatric disorders, mood, anxiety, prevalence rates and indeed may exaggerate the and health-related anxiety symptoms, and distrust role of risk factors, such as mental health history.
in authorities (alienation) are not only elevated To date, only three epidemiologic disaster studies (e.g. 4–5) but also often become intractable, of random community samples and adequate especially after human-made catastrophes (6–8).
controls (demographically similar unexposed pop- Recent disasters, such as the September 11 World ulations) have used a pre–post design: the Epi- Trade Center catastrophe, have also had a psy- demiologic Catchment Area follow-up after a set chological impact on indirectly exposed popula- of human and natural disasters (12); the Puerto tions (through television primarily) although the Rico follow-up of victims of a mud-slide (13, 14); strategies for configuring these samples raise seri- and the Netherlands follow-up of high school ous concerns about generalizability (e.g. 9, 10). In students who survived a cafe´ fire (15). These studies spite of the uniqueness of each disaster in terms of found increases in psychosomatic and post-trau- severity, sociocultural context, and postdisaster matic stress disorder (PTSD) symptoms, but not response and of differences in methodologies to evaluate the psychiatric consequences, some high- The present study adds significantly to this small risk groups have been consistently identified, body of pre–post studies by investigating a com- including mothers of young children, individuals munity sample before and after an air show Psychological aftermath of an air show disaster At the site of disaster Knew someone at the site Postdisaster risk factors
Watched television repeatedly
Perceived threat
Perceived influence on life
Fig. 1. Graphical representation of the Emotional support disaster in Ukraine in 2002 that killed or seriously airport's tarmac littered with body parts and injured over 100 people. Figure 1 shows the human remains. Stories about the plane crash – conceptual framework for the study. Because the worst of its kind in 14 years – appeared on the predisaster baseline data were available, we were front pages of newspapers worldwide. In Septem- able to examine mental health risk factors that ber 2002, EJB and JMH were in Lviv to discuss were free of disaster-associated recall bias.
the recently completed field work in Ukraine forthe World Mental Health (WMH) prevalencestudy of psychiatric and substance disorders Aims of the study (16), and were impressed by the psychological The aim of the study was to examine the mental turmoil the disaster appeared to have unleashed health effects of a gruesome air show disaster in on the community. The psychiatrists participating Lviv, Ukraine, in July 2002, in a population who in the WMH study were actively involved in previously participated in an epidemiologic mental providing mental health support to the stricken health survey. We examined differences in post- community, having translated American materials traumatic stress symptoms, somatization symp- on PTSD and provision of care after September toms, and anomie between the Lviv group and 11 that they disseminated to professionals and controls from western Ukraine and the relation- community members. The present study resulted ships of pre- and postdisaster risk factors to these directly from these discussions.
mental health outcomes.
Material and methods In 2002, we conducted a national survey of mentalillness and substance disorders in Ukraine (17) as part of the WMH initiative (18). The Ukraine On July 27, 2002, a 16-ton SU-27 warplane WMH study is a nationally representative survey smashed into a crowd of air show spectators at of residents aged 18 and older from the 24 oblasts Skniliff airbase outside Lviv (population 830 000) (counties) and the autonomous republic of Crimea in western Ukraine, after failing to recover from a (for details, see 17). Briefly, face-to-face interviews steep, low-altitude turn in which the pilot lost were carried out with 4725 respondents by the control of his plane's trajectory. The pilot sur- professional interview field staff of the Kiev Inter- vived by ejecting himself from the cockpit, but the national Institute of Sociology (KIIS) in collabor- plane crashed, killing 85 spectators (19 children) ation with the Ukrainian Psychiatric Association on the ground and injuring 151 adults and (UPA). Interviewers explained the study and children, 23 of whom seriously. During the obtained written informed consent prior to begin- hours and days that followed, the corridors of ning each interview. The recruitment, consent, and hospitals in the area were filled with relatives field procedures were approved by the Human looking for lost family members. Television crews Subjects Committees of University at Stony Brook, at the scene aired gruesome pictures of the KIIS, and UPA. The response rate was 78.3%.
Bromet et al.
The WMH fieldwork in western Ukraine, where 5 ¼ very much; Cronbach's alpha ¼ 0.93); soma- the accident occurred, was completed 1 month prior tization symptoms assessed with 12-item somati- to the disaster. There were 92 respondents (response zation subscale of the Symptom Checklist-90-R rate 89.2%) in the Lviv area where the accident (21) which rates symptoms over the past 2 weeks occurred, and 90 controls (response rate 88.7%) in on a 5-point severity scale (0 ¼ not at all; 4 ¼ two western regions of Ukraine, Rivne and Ivano- extremely; alpha ¼ 0.85); and anomie, a partic- Frankivsk, that served as the comparison site.
ularly significant concern in Eastern Europe (22), Follow-up face-to-face interviews were conducted assessed with a 4-item scale indicating whether in November–December 2002, 6 months after the respondents felt less safe, less able to control the disaster, with 75 Lviv respondents (81.5%) and 77 forces that influence their lives, more pessimistic about their future well-being, and had less faith in approved by Stony Brook and KIIS. There were the government's ability to protect them than no demographic differences in either site between before the air show disaster (1 ¼ not at all true; participants and non-participants in the follow-up 5 ¼ extremely true; alpha ¼ 0.78). Scoring for except that in Lviv, the participation rate was higher each scale involved summing across items.
in men (89.8%) than women (72.2%) (chi-square ¼ The follow-up interview also included a modified 4.77; df ¼ 1; P < 0.05).
version of the WMH-CIDI DSM-IV PTSDmodule that focused on the air show accidentand the WMH-CIDI modules for neurasthenia, Assessment and measures mood disorders, anxiety disorders, and heavy use The main tool of the initial survey was the WMH of alcohol, defined as either binge drinking (80 g of version of the Composite International Diagnostic pure ethanol a least once/month) or high frequency Interview for DSM-IV (WMH-CIDI), a fully of use (60 g 3–4 days/week or 40 g nearly every structured lay-administered diagnostic interview that generates DSM-IV and ICD-10 diagnoses(19). The CIDI was translated into Russian and Ukrainian using standard forward and back trans-lation procedures. Five variables from the initial Differences between the groups were analyzed survey were included as predisaster risk factors: using t-tests and odds ratios (95% confidence age, sex, education, financial adequacy, and life- intervals). Pearson correlations were used to exam- time disorder (DSM-IV mood, anxiety, and alco- ine bivariate associations. A series of linear regres- sion analyses was performed to estimate i) the Financial adequacy was categorized as ÔadequateÕ unique effects of exposure on symptomatology and if there was enough money for durables and anomie; ii) the group difference that remained after ÔinadequateÕ if there was not enough money for controlling for the predisaster risk factors; iii) the clothing or food.
group difference that remained after controlling for An Air Show Disaster Module was designed for the postdisaster risk factors; and iv) the group the follow-up study. The exposure section inquired difference that remained after controlling for var- about direct exposure (being at the event or iables that were statistically significant (P < 0.05) knowing someone involved), degree of perceived in the pre- and postdisaster regression analyses.
threat (based on the number of DSM-IV PTSD A1criteria endorsed, e.g. felt personally threatened, traumatized, terrified or very frightened at thetime, helpless, shocked or horrified, and numb), Compared with controls, the Lviv sample was whether respondents watched the television cover- significantly older (mean ± SD: 49.7 ± 18.6 vs.
age of the event repeatedly (vs. less often), whether 43.6 ± 14.1), included more men, and was more respondents believed that the disaster had an likely to report financial problems (Table 1). The important influence on their lives (vs. little or no groups were not significantly different on educa- influence relative to other events), and emotional tional attainment and predisaster lifetime psychi- support (number of types of people respondents atric disorder although the rate was somewhat turned to for support ÔsomeÕ or Ôa lotÕ as a result of higher in the Lviv group (33.3%) than in the the disaster). Postdisaster mental health included controls (22.1%).
PTSD symptoms assessed with the 22-item Impact Twenty-three respondents (including one con- of Events Scale-R (20) which rated severity of trol) knew someone who was killed or injured at intrusion, avoidance, and hyperarousal symptoms the accident. Two Lviv residents were at the resulting from the air show disaster (1 ¼ not at all; airbase. As expected, a larger proportion of the Psychological aftermath of an air show disaster Table 1. Distribution of key variables by site Lviv (n ¼ 75) % Controls (n ¼ 77) % Predisaster risk factors 2.11 (1.10–4.02)* Age (>45 years) 0.45 (0.23–0.86)* Education, £high school 1.17 (0.60–2.26) Finances, inadequate 7.35 (2.98–8.09)*** Mental or substance disorder 0.77 (0.86–3.63) Disaster risk factor Attended event or knew someone killed/injured 31.5 (4.1–241.3)*** Postdisaster risk factors Watched television repeatedly 5.42 (2.43–12.09)*** Perceived influence on life 2.30 (0.87–6.06) Emotional support Outcome variables Somatization symptoms *P < 0.05, **P < 0.01, ***P < 0.001.
Lviv sample watched the television coverage of the Lviv, met criteria for heavy alcohol use during the event over and over again and reported feeling threatened or horrified by the event (Table 1). Thegroups were not significantly different with respect Bivariate relationships between the risk factors and to emotional support received after the event or the extent to which they felt that the event had asignificant influence on their lives (although twice A number of relationships were examined (three as many Lviv residents endorsed this item).
variables · nine groups). We thus focus on correlations whereP < 0.01. Among the predisaster risk factors, only Differences in postdisaster mental health lifetime mental/substance disorder was strongly With respect to symptom severity, compared with related to the outcome measures. However, the controls, the Lviv sample had significantly higher postdisaster risk factors of perceived influence of PTSD symptom scores, but similar levels of soma- the event and perceived threat were strongly tization (Table 1). We note, however, that the associated to both PTSD symptoms and anomie single item from this scale which corresponds most in both groups. Repeatedly watching the disaster closely to the Ukrainian and Russian idiom of on television and social support were only weakly distress, i.e. Ôdo you feel heart pain,Õ was signifi- related to the outcome measures. Moreover, 1.87 ± 0.96 vs. 1.57 ± 0.70 for controls; t ¼ combined the two sites, repeatedly watching P < 0.05).
television was significantly related to PTSD symp- reported significantly greater anomie than the toms (P < 0.05). In the Lviv sample, being at the Lviv sample.
event or knowing someone who was injured or On a diagnostic level, 11 respondents from Lviv killed was not significantly related to the outcomes (14.7%) compared with no controls had episodes although it was correlated with perceived threat of depression, anxiety, neurasthenia, and/or PTSD (r ¼ 0.36; P < 0.01) and emotional support (r ¼ during the 6 months following the disaster (chi- 0.30; P < 0.01).
square ¼ 12.2; P < 0.001). Six of 11 with disorder(54.5%) knew someone who was killed or injured Multivariate analyses compared with 16 of 64 without postdisasterdisorder (25.0%; Fisher's exact test ¼ n.s.). Four The final analysis considered the ability of the pre- of 11 cases experienced their first lifetime episode and postdisaster risk factors to account for group during this time, corresponding to an incidence differences in the outcome measures. Consistent rate of 8.0% (4/50 with no lifetime disorder at with Table 2, the unadjusted regression coefficients wave 1). Two respondents knew someone killed or for group (row 1 of Table 3) were significant injured, and two did not. One respondent, from for PTSD and anomie. A comparison of these Bromet et al.
Table 2. Relationship of background and disaster characteristics to mental health: Pearson correlationcoefficients Predisaster risk factors Inadequate finances Mental or substance disorder Disaster risk factor Attended event or knew someone killed/injured )0.00 Postdisaster risk factors Watched television repeatedly Perceived influence on life *P < 0.05, **P < 0.01, ***P < 0.001.
Table 3. Accounting for differences between the Lviv and control groups before and after adjusting for pre- and postdisaster risk factors Group differences Adjusted for predisaster risk factors  Adjusted for postdisaster risk factors  Adjusted for all significant risk factors à  Unstandardized regression coefficients.
àAdjusted for mental health history, perceived threat, perceived influence on life, and support.
*P < 0.05, ***P < 0.001.
unadjusted coefficients with the adjusted coeffi- analysis of PTSD and somatization symptoms, but cients that controlled for the predisaster risk remained highly significant in the analysis of factors (row 2 of Table 3) showed that the group anomie. Specifically, in the final model for PTSD, difference in PTSD symptom severity was no the significant predictors were lifetime disorder longer statistically significant, while that for (P < 0.02), perceived threat (P < 0.001), and anomie, although reduced in size, was still highly influence on life (P < 0.02), but not exposure significant. The unadjusted and adjusted regression group. The significant predictors of somatization coefficients for somatization were non-significant.
were lifetime disorder (P < 0.01), perceived threat We note that the only predisaster variable that was (P < 0.01), and emotional support (P < 0.05), but significantly related to the outcomes in the regres- not exposure group. In contrast, the significant sion analyses was lifetime disorder (P < 0.01 for predictors of anomie included lifetime psycho- all three outcome variables; data not shown).
When the multivariate analyses were repeated (P < 0.001), emotional support (P < 0.001), as controlling for the postdisaster risk factors (row 3 well as exposure group (P < 0.001; controls had of Table 3), the group difference in anomie greater anomie). We also examined the interaction remained highly significant although it was again coefficients for the various pairs of risk factors, and reduced substantially compared with the unadjust- none was significant.
ed coefficient. The adjusted coefficients for PTSD Table 3 also shows that for each outcome, the and somatization were not significant. Except for final adjusted model explained substantially more repeatedly watching the television coverage, the of the variance than the unadjusted model. For other postdisaster risk factors were all significant in PTSD symptoms, the percentage increased nine- most of the models.
fold. For anomie, the percentage of variance The final multivariate models thus adjusted for explained doubled. This was due mostly to the mental health history, perceived threat, perceived postdisaster risk factors. For somatization, the influence of the event, and repeatedly watching increase was also substantial, television. As shown in Table 3 (row 4), the regres- to mental health history assessed before the sion coefficients for group were not significant in the Psychological aftermath of an air show disaster the event did not threaten the physical health of the general community (26).
This study represents one of the rare opportunities The most important predisaster risk factor was in which baseline data had been collected prior to having a history of mental disorder, as has been the occurrence of a community-wide disaster. To shown in several other studies (11, 25, 27). The fact our knowledge, it is the first pre–post study to be that this remained significant in the final model conducted in a former Soviet Union setting.
confirms that when identifying at-risk cases after a Previous research after the Chornobyl disaster in disaster, e.g. for prevention or early intervention, Belarus (6, 23), Ukraine (7), and Russia (8) has such individuals should be a primary target, even shown that is quite feasible to achieve high in non-Western settings. Age was not a risk factor standards of scientific rigor and reliability of for postdisaster symptomatology although in pre- standardized assessments in these settings.
In this study we demonstrated that the Lviv air Although gender and financial adequacy were show disaster had a considerable effect on the significant risk factors for postdisaster mental severity of post-traumatic stress symptoms in the health in many previous studies (1), they too Lviv sample and on diagnosable psychopathology.
were not significant in this study.
The unexpected finding that the degree of anomie In contrast, feeling threatened by the event and was lower in the Lviv sample than in the control perceiving it as an occurrence that significantly group may have resulted from the increased influenced one's life were the two most significant attention given to Lviv after the accident. Two postdisaster risk factors. This supports the increas- decades ago, Quarantelli (24) postulated that ing awareness of the importance of risk perception disasters may have positive (e.g. on social coher- and hazard perception as moderators of postdis- ence) as well as negative effects, but to our aster psychopathology (e.g. 27, 28), although one knowledge, no earlier study has actually documen- might argue that perceiving an event as a threat is ted any evidence to support this. Thus, our finding almost tantamount to an anxiety-related outcome on anomie, while extremely interesting, needs to be variable, rather than being a risk factor.
confirmed by future disaster research.
An inherent weakness of our study, and indeed Even though the Lviv sample was mostly all pre–post studies to date, is the modest sample composed of secondary disaster victims, i.e.
size. This was particularly relevant to our analysis people indirectly exposed, the postdisaster preval- of diagnosable disorders for which we combined ence rate was close to 15% and the postdisaster several disorders and could not examine disaster incidence rate was 8%. Although not statistically risk factors. A second weakness was that follow-up significant, the cases were mostly concentrated in interview took place 6 months after the disaster, the subgroup of primary disaster victims. These i.e. too late to capture immediate stress reaction in results were remarkably similar to our findings for the population. Other weaknesses include lack of mothers of young children during the year dimensional symptom data at baseline, potential following the Three Mile Island accident, for recall bias of the postdisaster risk factors, and the whom the 1-year prevalence rate was 14% and the lack of information on resilience factors.
1-year incidence rate was 11% (25). More import- In conclusion, the community surrounding the antly, this finding confirms prospectively that the Lviv air show disaster had more severe PTSD disaster had an influence on diagnosable disorder, symptomatology 6 months after the event, but the which previous prospective reports found to be a controls reported a greater sense of anomie in the trend (12, 13).
wake of the disaster. The key risk factors were In contrast to the findings of Escobar et al. (14), having a history of mental illness or substance we did not find a significant difference in somatic abuse, perceiving the event as threatening, and symptoms. We note that the Lviv group did have perceiving it as having a major influence on one's slightly higher scores than the controls, but both life. Our findings in western Ukraine confirm that groups were elevated compared with Western persons with a history of mental disorder should be samples (21). However, we did find a significant a primary target for research on the effectiveness of difference for the item which most appropriately early interventions. In light of the debate about the expresses the local idiom of distress (Ôheart painÕ).
usefulness of early interventions, future research on The most parsimonious explanations for our fail- the effectiveness of such interventions should ure to detect a significant difference in somatic perhaps focus on this high risk group. With respect symptom severity overall are the higher base rates, to the prevention of mental health consequences of the modest sample size, and perhaps more import- disasters, our findings suggest that the influence of antly, the fact that, in contrast to toxic disasters, the media may not be readily generalizable across Bromet et al.
events and cultural contexts. In the current age of 11. Bromet EJ, Dew MA. A review of psychiatric epidemiologic terrorist threat, it may be particularly relevant to research on disasters. Epidemiol Rev 1995;17:113–119.
seek new ways to mitigate the wider sense of 12. Robins LN, Fischbach RL, Smith EM et al. Impact of disaster on previously assessed mental health. In: Shore JH, ed.
insecurity and anomie that events like these can Disaster stress studies: new methods and findings. Wash- have on the population at large by focusing more ington, DC: American Psychiatric Press, 1986:21–48.
broadly on people's perceptions of these events in 13. Canino G, Bravo M, Rubio-Stipec M et al. The impact of addition to addressing their symptomatology.
disaster on mental health: prospective and retrospectiveanalyses. Int J Ment Health 1990;19:51–69.
14. Escobar JI, Canino G, Rubio-Stipec M et al. Somatic symp- toms after a natural disaster: a prospective study. Am JPsychiatry 1992;149:965–967.
Supported by the U.S. National Institute of Mental Health 15. Reijneveld SA, Crone MR, Verhulst FC et al. The effect of a (MH61905). The baseline survey was conducted as part of the severe disaster on the mental health of adolescents: a WHO World Mental Health survey consortium directed by controlled study. Lancet 2003;362:691–696.
Ronald C. Kessler, PhD (Harvard University) and T. Bedirhan 16. The WHO World Mental Health Survey Consortium.
Ustun, MD (World Health Organization). We thank psychi- Prevalence, severity, and unmet need for treatment of atrists Yuriy Zakal, Sergey Mykhnyak, Oksana Kulik, Roman mental disorders in the World Health Organization World Boguta, and Dmitriy Kanavin from Lviv, Ukraine, for sharing Mental Health Surveys. JAMA 2004;291:2581–2590.
their clinical experiences following the disaster; Vladimir 17. Bromet EJ, Gluzman SF, Paniotto VI et al. Epidemiology of Paniotto, Valeriy Khmelko and Victoria Zakhozha (Kiev psychiatric and alcohol disorders in Ukraine: findings from International Institute of Sociology) and Julia Pievskaya the Ukraine World Mental Health Survey. Soc Psychiatry (Ukrainian Psychiatric Association) for their roles in conduct- Psychiatr Epidemiol, in press.
ing the study; and Charles PM Webb for overseeing all aspects 18. WHO World Mental Health Survey Consortium. Preval- of the study. Special thanks go to the project participants for ence, severity, and unmet need for treatment of mental generously giving us their time and insights.
disorders in the World Health Organization World MentalHealth Surveys. JAMA 2004;291:2581–2590.
19. Kessler RC, U ¨ n TB. The World Mental Health (WMH) Survey Initiative Version of the World Health Organiza-tion (WHO) Composite International Diagnostic Interview 1. Norris FH, Friedman MJ, Watson PJ et al. 60,000 disaster (CIDI). Int J Methods Psychiatr Res 2004;13:93–121.
victims speak: Part I. An empirical review of the empirical 20. Weiss DS, Marmar DR. The impact of event scale – revised.
literature, 1981–2001. Psychiatry 2002;65:207–239.
In: Wilson JP, Keane TM, eds. Assessing psychological 2. WeisÆth L. Disasters: psychological and psychiatric trauma and PTSD. New York: Guilford Press, 1997:399– aspects. In: Goldberger L, Breznitz S, eds. Handbook of stress, 2nd edn. New York: Free Press, 1993:591–616.
21. Derogatis LR. Symptom Checklist-90-Revised: adminis- 3. Catapano F, Malafronte R, Lepre F et al. Psychological tration, scoring, and procedures manual – II. Towson, consequences of the 1998 landslide in Sarno, Italy: a MD: Clinical Psychometric Research, 1982.
community study. Acta Psychiatr Scand 2001;204:438– 22. Genov N. Global trends and Eastern European societal transformation. Int Soc Sci J 2000;52:539–547.
4. Dohrenwend BP. Psychological implications of nuclear 23. Havenaar JM, Rumyantzeva GM, Poelijoe NW et al. The accidents: the case of Three Mile Island. Bull N Y Acad reliability and validity of two psychiatric screening ques- tionnaires in the Russian Federation. Int J Methods Psy- 5. Kilic C, Ulusoy M. Psychological effects of the November chiatr Res 1996;6:237–242.
1999 earthquake in Turkey: an epidemiological study. Acta 24. Quarantelli EL. What is a disaster? The need for clarification Psychiatr Scand 2003;108:232–238.
in definition and conceptualization in research. In: Sowder 6. Havenaar JM, Rumyantseva GM, Van Den Brink W et al.
BJ, ed. Disasters and mental health: selected contemporary Long-term mental health effects of the Chernobyl disaster: perspectives. Rockville, MD: NIMH, 1984:41–73.
an epidemiological survey in two former Soviet Regions.
25. Bromet EJ, Parkinson D, Schulberg HC et al. Mental health Am J Psychiatry 1997;54:1605–1607.
of residents near the TMI reactor: a comparative study of 7. Bromet EJ, Goldgaber D, Carlson G et al. Children's well- selected groups. J Prev Psychiatry 1982;1:225–275.
being 11 years after the Chornobyl catastrophe. Arch Gen 26. Bromet EJ, Havenaar JM. Mental health consequences of disasters. In: Sartorius N, Gaebel W, Lo´pez-Ibor JJ, Maj M, 8. Viinama¨ki H, Kumpusalo E, Myllykangas M et al. The eds. Psychiatry in society. London: John Wiley & Sons, Chernobyl accident and mental wellbeing – a population study. Acta Psychiatr Scand 1995;91:396–401.
27. Bergiannaki JD, Psarros C, Varsou E et al. Protracted acute 9. Silver RC, Holman EA, Mcintosh DN, Poulin M, Gil-Rivas stress reaction following an earthquake. Acta Psychiatr responses to September 11. JAMA 2002;288:1235–1244.
28. Havenaar JM, de Wilde EJ, Van Den Bout J et al. Perception 10. Ahern J, Galea S, Resnick H et al. Television images and of risk and subjective health among victims of the Cher- psychological symptoms after the September 11 terrorist nobyl disaster. Soc Sci Med 2003;56:569–572.
attacks. Psychiatry 2002;65:289–300.

Source: http://upa-psychiatry.org.ua/articles/LvivActa.pdf