|
Health Problems of Victims Before and after Disaster: a Longitudinal Study in General Practice
International Journal of Epidemiology
By C Joris Yzermans, Gé A Donker*, Jan J Kerssens, Anja JE Dirkzwager, Rik JH Soeteman and Petra MH ten Veen
Advance Access published April 28, 2005
© The Author 2005; all rights reserved. doi:10.1093/ije/dyi096
NIVEL Netherlands Institute for Health Services Research, Utrecht, The Netherlands
* Corresponding author. NIVEL, PO Box 1568, 3500 BN Utrecht, The Netherlands. E-mail: g.donker@ nivel.nl
KEY MESSAGES
* Two and a half years post-disaster the prevalence of psychological problems in victims, who had to relocate, was about double and in the non-relocated victims one-third more than controls.
* Victims with pre-disaster psychological problems were at greater risk for post-disaster psychological problems.
* Relocated victims showed an excess of medically unexplained physical symptoms (MUPS) especially in a period of increased media attention.
Background
We aimed to quantify the health problems and to assess the possible risk factors for developing health problems in persons affected by the explosion of a firework depot at Enschede, The Netherlands, on May 13, 2000. The explosion considerably damaged buildings in the local neighbourhood and caused 22 immediate deaths and injuries in over 1000 people.
Methods
A longitudinal study of (89% of all) victims (n = 9329) and controls (n = 7392) with pre-disaster baseline morbidity for 16 months and post-disaster data for 2.5 years was conducted using the electronic medical records of general practitioners. Symptoms and diagnoses were recorded using the International Classification of Primary Care (ICPC). Prevalence rates for clusters of symptoms were compared between victim and control groups pre- and post-disaster. Risk factors for developing health problems were examined in hierarchical linear models.
Results
Two and a half years post-disaster, the prevalence of psychological problems in victims who had to relocate was about double and in the non-relocated victims one-third more than controls. Victims with pre-disaster psychological problems were at a greater risk for post-disaster psychological problems. Relocated victims showed an excess of medically unexplained physical symptoms (MUPS) especially in a period of increased media attention. Both groups of victims showed some increase of gastrointestinal (GI) morbidity 2.5 years post-disaster compared with their pre-disaster rate, and compared with the control group. Conclusions Two and a half years post-disaster an excess of psychological problems, MUPS, and gastrointestinal morbidity was observed. Pre-disaster psychological problems and inevitable relocation were predictors of more post-disaster psychological problems.
Keywords
Disaster, general practice, longitudinal, health problems, psychological, baseline
Many studies suggest that disasters and other traumatic events have long-term health consequences, especially on the occurrence of psychiatric disorders including post-traumatic stress, major depression, and anxiety; on behavioural disorders such as those associated with substance abuse and domestic violence; and on functional disabilities.1-9 Evidence of adverse effects (short and long term) can be found in questionnaires measuring psychological distress.5-7,10-12 Increased self-reports of non-specific psychological and medically unexplained physical symptoms (MUPS) have been observed many years after civil and military traumatic events.1-3,7,12-14 Analysis of the effects of disasters on health present particular difficulties for researchers since they can only be studied retrospectively. There are very few longitudinal studies that compare effects before and after exposure.15-20 Some studies on rescue workers and victims of natural disasters have been published and these are mostly based on self-reported questionnaires.18-20 Studies of long-term health problems in which health care data collected before a disaster are compared with similar data after are rare.17,21 The study described in this article compares psychological and physical health problems in victims and controls before and after major explosions in a residential area.
On the 13th of May, 2000, a firework depot at Enschede, The Netherlands, exploded and destroyed a large part of a multicultural neighbourhood. Twenty-two people died and about 1000 were injured.22 In a wider area, the explosions inflicted damage to houses, resulting in an evacuation of over 1000 residents. Many known risk factors for post-disaster morbidity applied in this community including evacuation, low socioeconomic status (SES), low level of education, and immigrant status.1,11 The extent and impact of the disaster urged the Dutch government to offer support to the regional authorities. An information and advice center (IAC) capable of dealing with all related problems was established; a registration system of victims was created; long-term monitoring of health problems was initiated; and an integrated approach for dealing with the psychosocial consequences of the disaster was facilitated.22 The present study used electronic medical records (EMR) in primary care, in order to quantify changes in the health problems in the community which could be attributed to the disaster and to identify risk factors predisposing to them. We compared health problems presented to the general practitioners (GPs) in the period one year and four months before the explosion with that in the 2.5 years after it. The study describes the evolution of post-disaster health problems and tests the role of several predictors of health problems.
Methods
Design
Recording by GPs in EMRs established before the disaster was maintained throughout the study period and was used to monitor health problems. Victims on the lists of these GPs were identified and an age and gender matched group of controls was recruited from the same practices. In the Dutch health care system it is obligatory for each citizen to be registered on the list of one GP, who must first be consulted if referral to secondary care is needed. Medical specialists report the results of consultations and interventions back to the GP.
Data were extracted from the EMRs of GPs in Enschede from 1 year and 4 months before the disaster to 2.5 years after it. The pre-disaster period was chosen because electronic data registration was well implemented by most of the GPs in Enschede by December 1998. Data extraction commenced in October 2001 (1.5 year post-disaster). Registration systems in general practice document the patient's symptoms, examination findings, diagnosis (or evaluation), and interventions, in accordance with the International Classification of Primary Care (ICPC), which is compatible with ICD-10.23 This study is concerned with symptoms and diagnoses.
Subjects
All 60 GPs in the city of Enschede were invited to participate: 44 (73%) agreed to do so and 89% of the victims were registered in these practices. Patients were informed about the participation of their GP and none refused to participate. Data collection was performed in accordance with the privacy protection procedures of the Dutch Data Protection Authority.
Victims and controls
Victims were identified from self-reporting and the zip-code of their home address at the time of the disaster, and from the registration list of the IAC. The EMRs of victims were flagged to identify them easily. A control group matched by age and gender was recruited from patients on the lists of participating GPs. For analysis, victims were separated into two groups: those who had to relocate for 4 years owing to serious damage to their houses, and those who did not. Previous research after natural disasters had suggested that major loss of personal belongings increased the likelihood of subsequent health problems.14
Data analysis
Baseline demographic data of the two groups of victims and the controls were compared using Chi-square tests. Results focus on the total number of consultations and on consultations for diseases grouped in clusters including psychological problems, MUPS, and gastrointestinal (GI) morbidity. The choice for these clusters was based on published reports of increased prevalence of psychological problems and of MUPS, particularly GI-symptoms, after disasters.7,14,24,25 MUPS are somatic symptoms (tiredness, headache, nausea, and abdominal pain) without a diagnosis by the GP. Most musculoskeletal symptoms are included in the cluster MUPS and thus were not analysed as a separate cluster.24 The clusters are not exclusive, a patient could present symptoms in one or more clusters.
Prevalence rates in 4 week periods were calculated as the number of individuals presenting symptoms divided by the number of individuals at risk, taking into account the amount of person-time during which events were counted as well as the time elapsed before health symptoms were presented.26 Thus, the data were adjusted for loss to follow-up owing to death or moving to another area.
Series of linear regression models were tested with various independent time variables in order to analyse trends in the number of patients presenting problems to their GP before and after disaster. Time was measured in 49 periods of 4 weeks. The time variable 'Immediately after' applies to the first 4 weeks after the disaster and 'Trend after' is concerned with trends over the entire post-disaster study period. The term 'Level after' describes the comparison between pre- and post-disaster measurements.
Hierarchical linear models were applied in order to identify predictors of health problems in different groups of victims.27 Four outcome variables were analysed: whether or not a victim consulted the GP in a given period for any problem, for psychological problems, for MUPS, or for GI morbidity. In order to minimize the number of zero values in these models time was considered in 15 periods of 13 weeks (ending 8 November, 2002). Regression models have to be hierarchical as observations are repeated measurements of the same persons, so these are not independent. The variables 'Immediately after', 'Trend after' and 'Level after' were defined as in the linear regression analyses except for the difference in time period (13 weeks instead of 4 weeks). Gender, age, SES and various interaction variables between patient characteristics and time were computed. Age was computed every decade, centred around its mean. SES was defined on the basis of health insurance category (sick-fund vs privately insured; the higher income level is related to private health insurance). Data were analysed with 'Multilevel Models for Windows' (MlwiN).28
Results
Demographics and morbidity
A total of 16 721 patients were successfully monitored by the GPs for 3 years and 9 months. All victims on the lists of the participating GPs were included: 935 victims who had to relocate; 8394 non-relocated victims and 7392 controls (matched for gender and age to the group of victims, Table 1). In total, ~90 000 persons were on the lists of the 44 participating GPs. Both groups of victims were more likely to be publicly insured for health compared with the control group. One-third of relocated victims and one-fifth of the non-relocated victims were immigrants (no comparable data for the control group). The percentages of patients not presenting any problem to the GP during the study period were small in both groups of victims (together 6.1%) and in controls (7.7%) (Table 1).
Analyses of all symptoms and diagnoses reported to the GP showed a peak of contacts in both groups of victims immediately after the disaster with particularly high rates of contact in the first two periods of 4 weeks post-disaster. Predisaster, the consultation rate in the two groups of victims was higher than in the control group. During the 2.5 years post-disaster, differences in the overall consultation rate between victims and controls were observed, but rates recorded at the end of the study in both the victim groups and the control group were almost the same as the pre-disaster equivalents.
The pre-disaster prevalence of psychological problems in both groups of victims was similar to that in controls; immediately after the disaster there were very substantial increases in both victim groups (up to a 15-fold increase during the first month post-disaster among the relocated victims), while the prevalence of psychological problems in the control group was constant throughout (Figure 1). Although prevalence rates for psychological problems decreased gradually in both victim groups, by the end of the study the rates were still substantially higher compared with pre-disaster and with the control group (for the relocated victim group twice the prevalence rate of controls and for the non-relocated victims a 31% higher prevalence rate compared with controls).
Victims and controls with pre-disaster psychological problems showed a 4- to 7-fold prevalence of psychological problems post-disaster compared with those without pre-disaster psychological problems. The highest prevalence rate of psychological problems immediately after the disaster and throughout the study was observed in those who had to relocate and already presented psychological problems pre-disaster (750 per 1000 in the first 4 weeks post-disaster).
Both groups of victims presented MUPS more often to the GP pre-disaster than the control group (Figure 2). The relocated victims reported a higher prevalence of MUPS immediately after the disaster (104 per 1000 in 4 weeks), and the prevalence continued at a higher level to the end of the study (88 per 1000 per 4 weeks). A high peak of MUPS in relocated victims was observed in April 2002 after the two managers of the firework depot were acquitted from all charges. This event received substantial media attention. The prevalence of MUPS in nonrelocated victims was constant throughout the whole period (76 per 1000 in 4 weeks). Of the symptoms included in MUPS cough, tiredness, and pain of neck, back, and shoulder were those most frequently reported in all three study groups. GI morbidity was more often presented to the GP by relocated victims pre-disaster and throughout the whole period (23 vs 18 per 1000 per 4 weeks in the control group). A small increase in GI morbidity was observed in both disaster groups 2.5 years post-disaster compared with both the control group and the pre-disaster rate.
Predictors of psychological distress, MUPS, and GI morbidity
The hierarchical linear modelling shows the high level of postdisaster psychological problems in both victim groups, but especially in the relocated victims (OR = 5.81; 95% CI 4.63-7.29). Female gender, older age, and low/middle SES were predictors of increased psychological problems in victim and control groups (Table 2), though (except for female gender in the immediate post-disaster period) they were not stronger predictors in the victim groups than in the control group.
The prevalence of MUPS, initially low in both groups of victims, shows an increasing trend post-disaster, especially in relocated victims (OR = 1.04; 95% CI 1.01-1.08). Female gender, increasing age, and low to middle SES were predictors for increased MUPS in the control group. Female gender did not predict more MUPS in both groups of victims while age (in nonrelocated victims only) and low to middle SES did. However, there were several interaction terms between the different groups of victims, age, gender, and SES (Table 2).
Male gender, increasing age, and low to middle SES did predict more GI morbidity in the control group while age and gender did not in any of the victim groups. Low to middle SES did predict more GI morbidity in the non-relocated victims compared with controls.
When assessing different hierarchical linear models including interaction terms: female gender, older age, and low SES appeared to be equal predictors of symptoms in the control group and in both groups of victims.
Ethnicity (not known for the control group) did not predict the number of psychological problems presented to the GP in both groups of victims pre-disaster (not in table). Separate analysis of immigrants by origin (Europe or non-Europe) revealed higher prevalence of psychological problems postdisaster in relocated non-European immigrants compared with the Dutch reference category (OR = 7.71; 95% CI 4.87-12.19). In the non-relocated immigrants long-term differences by ethnicity were not observed.
Discussion
This study offers an important contribution to disaster research as both pre- and post-disaster data could be analysed in victims and controls. More than 90% of persons consulted their GP during the observation period and provided the opportunity to gather a comprehensive dataset unaffected by biased victim recall. The study shows the vulnerability of victims who need to be rehoused for several years, and of victims previously experiencing psychological problems. The increase in problems presented to the GP by victims was sustained over 2.5 years following the explosion. The increases were noted especially for psychological problems, compared with both pre-disaster equivalents and with prevalence in the control group. Victims with psychological problems before the disaster were more at risk especially where they had experienced severe disruption to their lives needing to be relocated. Previous studies have shown that prior psychological problems are a strong predictor of further difficulties after disasters.1,15,29 However, these studies were mostly based on selfreported retrospective data: the conclusions of our study are based on a comparison of data available in EMR, which preceded the disaster with data collected after it observing the same recording protocol. Relocation and major loss of personal belongings have been recognized as important predictors of increased psychological problems after natural disasters.15,30,31Financial compensation was rounded off within 2 years after the disaster. Although unsettled compensation claims are well-known to increase the persistence of some psychological symptoms, we do not consider these to have a major influence on morbidity and health care utilization in our study owing to the government's generosity and the relatively fast and transparent procedures. The temporarily increased prevalence of MUPS in relocated victims during a period of increased media attention as showed in this study has not been demonstrated so clearly before.9
In our study women, the elderly, and those with lower SES reported more symptoms, but the higher susceptibility was similar in the victim and control groups. Earlier studies implicating female gender, lower SES, and older age as risk factors for post-traumatic stress disorder and psychological distress, were not based on study designs that included predisaster data and control groups.32-35
The study revealed differences in the prevalence of MUPS and GI morbidity with slightly higher rates in the victim groups present even before the explosion. Differences in SES and the percentage of victims who were immigrants may have contributed to the pre-disaster differences in disease prevalence. Both groups of victims, but especially the relocated victims, had a lower SES than the control group. Disasters occur more frequently in socially deprived areas.8 Immigrants have often experienced multiple traumatic events and sometimes difficulty in establishing themselves in a new environment: both of which may have caused psychological problems and perhaps explain the increased prevalence in pre-disaster morbidity. Limited availability of information about immigrant status prevented more detailed analysis. In our study non-European immigrant status predisposed to psychological problems post-disaster, which is consistent with the findings of others.5,36 Ethnic differences in coping with stress and depressive symptoms have also been reported after other disasters and may reflect a lack of social support, but also differences in the perception of symptoms.3,11,37
Some methodological issues must be considered. This study included victims either by self-classification or GP identification. A risk of overrepresentation of victims with symptoms would occur if GP registration had been taken as the sole source. Selfselection may include people minimally affected and seeking financial compensation could lead to an underestimation of the effects of a disaster unless there is an objective assessment of the adverse effects sustained.
A limitation of the study may be that the outcome variable is based on the GP evaluation of the problem. Though we recognize that the classification of a problem is the product of the patient-doctor interaction, we believe it to be a reliable assessment, which is more likely to underestimate the effects of the disaster. A direct influence of the disaster on GPs' interpretation of symptoms and a certain lack of completeness could not be excluded. This is a limitation of all studies on monitoring and registration. However, it offers the unique opportunity of comparison pre- and post-event.38 Morbidity surveys in general practice cannot be compared directly with data obtained from health interview surveys that measure patients' experience of health status.39
In our study, recorder bias was minimized by matching victims and controls by GP. Research on the World Trade Center terrorist attack has shown that the psychological effects were not limited to those directly exposed. The degree of response was not predicted simply by objective measures of exposure to or loss from the trauma.21,40 This is certainly the case for residents in the same city who experienced the disaster from close by and, most probably, observed the disaster site. If this affected the outcome of our study, the effects would be underestimated. However, it is unlikely that this effect would be sustained over 2.5 years.
Considering these limitations, the present study has some major strengths: the use of a control group and comparison of health data before and after disaster in victims and controls. The finding that relocation and pre-disaster psychological problems are determinants of increased problems post-disaster lasting several years, implies a need for long-term support of victims. Adequate recording of major life events and episodes in medical records is of utmost importance to support intervention strategies. This study demonstrates the long-term impact of the disaster on health and the value of organized data collection by EMR in general practice. Health professionals are made aware that victims may struggle with increased physical and psychological problems several years after a disaster.
Acknowledgments
We gratefully acknowledge the contribution of the participating general practitioners. This study was funded by a grant obtained from the Ministry of Public Health, Welfare and Sports in The Netherlands. We thank Dr Douglas Fleming of the Birmingham Research Unit of the Royal College of General Practitioners, for valuable comments on an earlier version of this article.
References
1 Dew MA, Bromet EJ. Predictors of temporal patterns of psychiatric distress during 10 years following the nuclear accident at Three Mile Island. Soc Psychiatry Psychiatr Epidemiol 1993;28:49-55.
2 Ismail K, Everitt B, Blatchley N et al. Is there a Gulf War syndrome? Lancet 1999;353:179-82.
3 Green BL, Lindy JD, Grace MC et al. Buffalo Creek survivors in the second decade: stability of stress symptoms. Am J Orthopsychiatry 1990;60:43-54.
4 Bland SH, O'Leary ES, Farinaro E, Jossa F, Trevisan M. Long-term psychological effects of natural disasters. Psychosom Med 1996;58:18-24.
5Winje D. Long-term outcome of trauma in adults: the psychological impact of a fatal bus accident. J Consult Clin Psychol 1996;64:1037-43.
6Marmar CR, Weiss DS, Metzler TJ, Delucchi KL, Best S, Wentworth KA. Longitudinal course and predictors of continuing distress following critical incident exposure in emergency services personnel. J Nerv Ment Dis 1999;187:15-22.
7 Donker GA, Yzermans CJ, Spreeuwenberg P, van der Zee J. Symptom attribution after a plane crash: comparison between self-reported symptoms and GP records. Br J Gen Pract 2002;52:917-22.
8 Rubonis AV, Bickman L. Psychological impairment in the wake of disaster: the disaster-psychopathology relationship. Psychol Bull 1991;109:384-99.
9 Yzermans CJ, Donker GA, Vasterman P. The impact of disasters: long term effects on health. In: Kirch W (ed). Public Health in Europe. 10 years EUPHA. Berlin: Springer, 2003, pp. 317-41.
10 Ursano R, Fullerton C, Kao T, Bhartiy V. Longitudinal assessment of posttraumatic stress disorder and depression after exposure to traumatic death. J Nerv Ment Dis 1995;183:36-42.
11 Yzermans J, Gersons BPR. The chaotic aftermath of an airplane crash in Amsterdan. A second disaster. In: Havenaar JM, Cwikel JG, Bromet EJ, eds. Toxic Turmoil. London, Kluwer Academic Publishers, 2002.
12 Ouimette P, Cronkite R, Henson BR, Prins A, Gima K, Moos RH. Posttraumatic stress disorder and health status among female and male medical patients. J Trauma Stress 2004;17:1-9.
13 Hyams KC, Wignall FS, Roswell R. War syndromes and their evaluation: from the US Civil War to the Persian Gulf War. Ann Intern Med 1996;125:398-405.
14 Escobar JI, Canino G, Rubio-Stipec M, Bravo M. Somatic symptoms after a natural disaster: a prospective study. Am J Psychiatry 1992;149:965-67.
15 Norris FH, Friedman MJ, Watson PJ, Byrne CM, Diaz E, Kaniasty K. 60 000 disaster victims speak: part I. An empirical review of the empirical literature, 1981-2001. Psychiatry 2002;65:207-39.
16 Norris FH, Friedman MJ, Watson PJ. 60 000 disaster victims speak: part II. Summary and implications of the disaster mental health research. Psychiatry 2002;65:240-60.
17 Reijneveld SA, Crone MR, Verhulst FC, Verloove-Vanhorick SP. The effect of a severe disaster on the mental health of adolescents: a controlled study. Lancet 2003;362:691-96.
18 Alexander D. Stress among police body handlers: a long-term followup. Br J Psychiatry 1993;163:806-08.
19 Nolen-Hoeksema S, Morrow JA. A prospective study of depression and posttraumatic stress symptoms after a natural disaster: The 1989 Loma Prieta earthquake. J Pers Soc Psychol 1991;61:115-21.
20 Dirkzwager AJE, Yzermans CJ, Kessels FJM. Psychological, musculoskeletal problems and sickness absence before and after involvement in a disaster: a longitudinal study among rescue workers. Occup Environ Med 2004;61:870-72.
21 Silver RC, Holman EA, McIntosh DN, Poulin M, Gil-Rivas V. Nationwide longitudinal study of psychological responses to September 11. JAMA 2002;288:1235-44.
22 Roorda J, van Stiphout WA, Huijsman-Rubingh R. Post-disaster health effects: strategies for investigation and data-collection. Experiences from the Enschede firework disaster. J Epidemiol Community Health 2004;58:982-87.
23 Lamberts H, Woods M. International Classification of Primary Care. Oxford: Oxford University Press, 1987.
24 Yzermans CJ, Dirkzwager AJE, den Ouden DJ et al. Monitoring gezondheid getroffenen Vuurwerkramp Enschede. Utrecht: NIVEL, 2003.
25 Schnurr PP, Spiro A, Paris AH. Physician-diagnosed medical disorders in relation to PTSD symptoms in older military veterans. Health Psychol 2000;19:91-97.
26 Rothman KJ, Greenland S. Modern epidemiology. Philadelphia: Lippincott Williams and Wilkins, 1998.
27 Snijders TAB, Bosker RJ. Multilevel Analysis. An Introduction to Advanced Multilevel Modelling. California: Sage Publications Ltd, 1999.
28 Rasbash J, Browne W, Goldstein H et al. A User's Guide to MlwiN. Multilvel Models Project, Institute of Education, University of London, London 2000.
29 North C, Nixon S, Shariat S et al. Psychiatric disorders among survivors of the Oklahoma City bombing. JAMA 1999;282:755-62.
30 Riad J, Norris F. The influence of relocation and the environmental, social and psychological stress experienced by disaster victims. Environ Behav 1996;28:163-82.
31 Najarian B, Goenjian A, Pelcovitz D, Mandel F, Najarian B. The effects of relocation after a natural disaster. J Trauma Stress 2001;14:511-26.
32 Foa EB, Street GP. Women and traumatic events. J Clin Psychiatry 2001;62 (suppl. 17):29-34.
33 Ticehurst S, Webster R, Carr V, Lewin T. The psychosocial impact of an earthquake on the elderly. Int J Geriatr Psychiatry 1996; 11:943-51.
34 Norris F, Kaniasty K, Conrad M, Inman G, Murphy A. Placing age differences in cultural context: a comparison of the effects of age on PTSD after disasters in the U.S., Mexico, and Poland. J Clin Geropsychol 2002;8:153-73.
35 Armenian H, Morikawa M, Melkonian A et al. Loss as a determinant of PTSD in a cohort of adult survivors of the 1988 earthquake in Armenia: implications for policy. Acta Psychiatr Scand 2000; 102:58-64.
36 Breslau N, Chilcoat HD, Kessler RC, Davis GC. Previous exposure to trauma and PTSD effects of subsequent trauma: results from the Detroit Area Survey of Trauma. Am J Psychiatry 1999;156: 902-07.
37 Palinkas LA, Russell J, Downs MA, Petterson JS. Ethnic differences in stress, coping, and depressive symptoms after the Exxon Valdez oil spill. J Nerv Ment Dis 1992;180:287-95.
38 Beale N, Nethercott S. Job-loss and family morbidity: a study of a factory closure. J R Coll Gen Pract 1985;35:510-14.
39 van der Velden J, Donker G, van der Steen J, van Sonsbeek JLA, van den Bos GAM. Methodology What do health interview surveys tell us about the prevalences of somatic chronic diseases? A study into concurrent validity. Eur J Public Health 1998;8:52-58.
40 Schuster MA, Stein BD, Jaycox LH et al. A national survey of stress reactions after the September 11, 2001, terrorist attacks. N Engl J Med 2001;345:1507-12.
|
