The European Journal of Public Health Advance Access originally published online on March 8, 2005
The European Journal of Public Health 2005 15(2):170-174; doi:10.1093/eurpub/cki126
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Public Mental Health |
Disaster and associated changes in physical and mental health in older residents
Dorly J.H. Deeg1,2, Anja C. Huizink1, Hannie C. Comijs2 and Tjabe Smid1
1 Institute for Research in Extramural Medicine, Vrije Universiteit Medical Center, Amsterdam, The Netherlands
2 Department of Psychiatry, Vrije Universiteit Medical Center, Amsterdam, The Netherlands
Correspondence: Prof. Dr D.J.H. Deeg, Vrije Universiteit Medical Centre, Institute for Research in Extramural Medicine/LASA, Van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands, tel. +31 20 444 6767, fax +31 20 444 6775, Email: djh.deeg{at}vumc.nl
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Background: Long-term health consequences of disasters have not been studied extensively, one reason amongst others is that no pre-disaster observation is available. This study focuses on an aeroplane crash on an Amsterdam suburb. The ongoing Longitudinal Aging Study Amsterdam has one pre-disaster and several post-disaster observations, making it possible to study changes in health, taking pre-disaster health characteristics into account. Methods: Three exposure groups are distinguished: those living within a radius of 1 km from the disaster (initial n=39), those living between a radius of 1 and 2 km from the disaster (initial n=56), and those living in the rest of the city of Amsterdam (initial n=508). Health measures include general health, health in comparison with age peers, functional limitations, disability and cognitive functioning. These measures are based on self-ratings, interviewer observations, or both. Results: Older persons living closest to the disaster area are likely to experience health decline in the wake of a disaster, over and above the health decline that would occur normally with aging. The disaster-associated health decline is small, and most obvious in the ability to perform actions (such as mobility), but is not observed in either disability in daily functioning, nor in self-perceptions of health. Cognitive functioning even shows a short-term improvement. Conclusion: These findings suggest substantial resilience in older adults, despite their common health problems.
Keywords: disaster, pre-post observation, health changes, older persons
On October 4, 1992, a cargo Boeing 747 crashed into two high rise apartment buildings in an Amsterdam suburb, the Bijlmer. Thirty-nine residents were killed, and many residents were directly or indirectly affected by the plane crash.
A man-made disaster such as the aeroplane crash in the Bijlmer may result in short-term and long-term health effects. Short-term effects include physical injuries due to the direct effects of exposure to the disaster, such as burns and fractures. These may result in long-term somatic complaints. Other short-term effects include psychological stress symptoms, such as intrusive thoughts of the traumatic experience and hyperarousal in the first weeks after the disaster.1 Long-term health effects of disasters include post-traumatic stress disorder,2,3 depression and more general health complaints.4,5 Complaints that are frequently reported after exposure to disasters or other stressful events include headache, fatigue, dizziness, concentrations problems and painful joints or muscles.6–8 Often, no somatic explanation is found for the reported complaints, and they are labelled as medically unexplained symptoms or functional somatic disorders.7,9,10 Such complaints are diffuse and non-specific, and they are commonly found in general populations.10,11
Health consequences of the exposure to a disaster have been studied mostly in retrospectively designed studies,4,5,12,13 or when part of the sample was interviewed before and part after the disaster.14 The availability of pre-crash and post-crash data from an on-going longitudinal study among older residents of the affected area in Amsterdam offered a unique opportunity to study changes in health in relation to the disaster. The hypothesis was tested that more negative changes in health over time are found in persons living closest to the crash area, when compared to persons living in areas at further distances from the crash site.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Sample
The Longitudinal Aging Study Amsterdam (LASA) is based on a nationally representative cohort, initial ages 55–85 years, with oversampling of men and older-old. The sample was recruited in 1992 (T0) for the Study on Living Arrangements and Social Networks of older adults (LSN), which had a response rate of 62.3% (n=3805).15 The current study is restricted to older persons living in Amsterdam at T0 (n=1066). For T0, all subjects were interviewed before the date of the disaster, October 4, 1992.
About 10 months after the LSN interview and at least 6 weeks after the disaster, the participants were approached for the first LASA cycle (1992–1993, T1).16 At this cycle, the Amsterdam cohort included 1028 surviving LSN participants (96.4%). Of the survivors, 865 subjects (84.1%) took part. By the second LASA cycle (1995–1996, T2), 13.9% of T1 participants had died. Of the 745 surviving participants, 699 (93.8%) were interviewed.17
Two analytic samples were defined for this study, to allow testing of both short-term and longer-term effects. The first (sample I) consists of those who participated at T0 and T1; the second (sample II), of those who participated at T0, T1 and T2.
The sample of participants was subdivided according to residential distance from the disaster as a proxy for exposure: (1) those living >2 km from the disaster (i.e. residents of Amsterdam exclusive of the Bijlmer suburb, n=971), (2) those living 1–2 km from the disaster (distant Bijlmer residents, n=56), and (3) those living <1 km from the disaster (near Bijlmer residents, n=39). Subsample (1) is used as the reference group.
Because the Bijlmer participants were significantly younger than those living in the rest of Amsterdam, the latter subsample was matched to the Bijlmer subsample on age, which procedure left 508 participants in this subsample.
Data
Distance to disaster was determined based on 6-digit postal codes. One apartment building is generally covered by several 6-digit postal codes.
Socio-demographics included sex and age, marital status (0=married, 1=never married, divorced and widowed), and socio-economic status (highest educational level attained and financial status based on the postal code) as measured at T0.
Physical and mental health indicators were measured at all three cycles.
Functional ability was assessed using six self-report items, e.g. climbing up and down a staircase and cutting one's own toenails. Five response categories ranged from ‘yes, without difficulty’ to ‘cannot’.18,19 The six items were combined into one score ranging from 6=unable to do all of the activities, to 30=having difficulty with none of the activities. In addition, the interviewer was asked to rate the respondent's mobility on a nine-point scale ranging from 1=very poor to 9=very good.
Disability was assessed using an item from the Medical Outcomes Study on health problems limiting daily activities, coded as: 1=no limitations, 2=mild limitations, 3=severe limitations.20
Self-rated health. A general question asking ‘How is your health in general?’ had five response categories ranging from 1=very good to 5=poor.21 A second question added ‘compared with your age peers’ with five response categories from 1=much better to 5=much poorer. In addition, after the end of each interview the interviewer was asked to rate the respondent's general health on a nine-point scale ranging from 1= very poor to 9=very good.
Mental health consisted of interviewer observations of the respondent's cognitive functioning during the interview, including seven aspects such as attention, comprehension, and memory. Ratings on these aspects were summed to a score ranging from 0=severe cognitive problems to 7=no cognitive problems.17
Statistical analysis
First, group differences in mortality and other sample attrition at T1 and T2, respectively, were tested in relation to distance from disaster. Second, differences in socio-demographic variables were tested among the three distance groups. Those socio-demographic variables that showed significant differences, were included as covariates in further analyses.
Third, differences in rates of change in physical and mental health were evaluated using multivariate analysis of variance for repeated measures. For the analyses of the short-term effects, the pre-disaster and one post-disaster assessments were included in the test of within-subjects effects (sample I). For the analyses of the longer-term effects, the pre-disaster and two post-disaster assessments were included (sample II).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Attrition
In the full study sample, the refusal rate at T1 was much greater than at T2, whereas the mortality rate increased gradually (table 1). 17 There were, however, differences among subsamples.
|
Non-mortality attrition appeared to be greater in the Bijlmer subsample nearest the disaster, but smaller in the distant Bijlmer subsample (relative risks of attrition compared to the reference sample at long-term follow-up: RR=1.77 and 0.47, respectively, P=0.064). In the full study sample, non-mortality attrition between T0 and T1 and between T0 and T2 was associated with poorer baseline health on all indicators except self-reported disability. These associations did not differ among subsamples, with one exception between T0 and T1; whereas in the reference sample, baseline self-reported functional ability was worse among those who refused, in the Bijlmer samples baseline self-reported functional ability did not differ between the refusers and the continuing participants.
Mortality did not differ significantly across the subsamples. In the full sample, mortality was associated with all baseline health indicators except self-reported disability. For all interviewer-rated indicators, however, the reference sample and the Bijlmer subsamples showed different associations: whereas the expected associations were observed in the reference sample, in the Bijlmer subsamples no associations between interviewer-rated health and mortality were observed.
These findings indicate a tendency for both mortality and non-mortality attrition to be less influenced by baseline health status in the Bijlmer residents than in the subjects living outside the Bijlmer suburb.
Baseline characteristics
The average initial age of the study sample was 68.2 years. Comparison of the subsamples with respect to age, education, income and marital status did not yield significant differences at P < 0.05 (table 2). The percentages of females and of unmarried respondents were somewhat smaller in the Bijlmer subsamples (P=0.058 and 0.087, respectively). The marital status difference disappeared when stratifying by sex, due to the much higher likelihood of women being unmarried at older ages. Because of the uneven distribution of sex in the subsamples, sex is included as a covariate in the analyses of health outcomes.
|
The health indicators showed few differences across the subsamples at baseline. Self-reported disability was slightly greater in the distant Bijlmer subsample (P=0.039). Self-reported general health was slightly worse in both Bijlmer subsamples (P=0.058).
Health outcomes
Significant declines in self-reported functional ability between T0 and T1 were demonstrated in sample I (table 3, main effect). The rates of decline were different for specific subsamles, as the interaction term between time, distance to disaster, and sex was significant. This indicates that females declined more than males, except in the subsample nearest to the disaster.
|
A significant decline in self-reported functional ability was also apparent over the period of four years, as demonstrated in sample II (table 3). Here, the interaction term between time and distance to disaster was significant, but a three-way interaction with sex no longer showed significance. This indicates that the longer-term rate of decline was greatest in the subsample nearest to the disaster, and that the longer-term rates of decline no longer differed by sex (figure 1).
|
There were no significant main or interaction effects for self-reported disability, either in sample I or in sample II (table 3).
Interviewer-rated mobility showed significant declines over time (table 3). Moreover, in sample I the interaction effect between time and distance to disaster was significant, indicating that the short-term decline was greater in the Bijlmer respondents. However, in sample II the longer-term differences in rates of decline did not reach significance (figure 2).
|
The respondents' self-rated health ‘in general’ showed a significant decline over time (table 3). However, since none of the interaction terms were significant, this decline was not different across the subgroups. This was true for both the short term and the longer term. Self-rated health in comparison with age peers showed no significant decline (table 3).
Interviewer-rated health showed a significant interaction effect of time and distance to disaster in sample I (table 3). As opposed to the Amsterdam respondents outside the Bijlmer suburb, the Bijlmer respondents showed short-term declines. In sample II, the main effect of decline was significant. However, the decline no longer differed significantly across the subsamples.
Cognitive functioning as rated by the interviewers showed no significant change over time in sample I (table 3, main effect). In sample II, the main effect of cognitive functioning was significant, indicating a decline over the four-year period (table 3). In both samples, however, the interaction term of time with distance to disaster was marginally significant (P=0.09). The mean scores showed a continuous decline in the reference sample. In contrast, in both Bijlmer subsamples, an initial increase was observed, followed by a decrease in cognitive functioning from T1 to T2 (figure 3).
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This study examined short-term and longer-term health effects of a disaster by making use of data from an ongoing longitudinal study with its initial data collection cycle a few months before the disaster, and subsequent data collection cycles a few months and several years after the disaster. As expected, more negative changes were observed on several indicators of health in those subjects who lived closest to the site of the disaster. This was true for self-reported functional ability, interviewer-rated mobility and interviewer-rated general health. The more negative changes in mobility and general health were only apparent shortly after the disaster. Self-reported disability and self-rated health did not show differences in rate of change among subjects living at shorter or longer distance from the site of the disaster. There were also no differences in mortality.
Other studies have suggested that disasters may result in various health effects.1–13 The findings of the present study suggest that the risk of negative health outcomes associated with a disaster is not pervasive across health indicators. There are a series of tentative explanations for this finding. First, functional ability, mobility and disability are concepts on one continuum, referred to as the disablement process.22 Limitations in functional ability, including mobility limitations, occur earlier in the process than disability, as functional limitations denote the inability to perform actions that are needed to carry out an activity. Such actions are not activities in themselves. Disability denotes restrictions in activities, in particular in role functioning. Functional limitations, then, may be a more sensitive measure of changes in health than disability. A more mundane reason for the lesser sensitivity of the disability measure that was used in this study, may be the shortness of its scale. With only three response categories, its sensitivity to change is likely to be limited.
Second, there is ample evidence for the relative stability of self-perceptions of health. Especially in older persons, who are often faced with inevitable health problems, self-rated health has been shown to be better than expected, based on objective health status.23,24 This relative stability is reflected by the barely significant longer-term decline in self-rated health in our sample. Interestingly, interviewer-rated general health did show longer-term decline, as well as greater short-term declines in the subsamples nearest to the site of the disaster. A stabilizing ‘trait’-component that may be at work in the respondents' self-ratings, is absent in the interviewer ratings. Therefore, the interviewer ratings may be more sensitive to changes over time. However, as the interviewers were aware that the Bijlmer respondents had experienced a recent disaster, they may have given a lower rating to these respondents' general health. This possibility of an existence of interviewer bias, however, is not confirmed by the interviewer ratings of cognitive functioning, that were higher shortly after the disaster. Moreover, the interviewers at T0 were not the same as at T1, and they did not know the T0 ratings.
Interviewer-rated cognitive functioning showed short-term improvement in those subjects living nearest to the site of the disaster. This unexpected finding deserves a special note. The interviewer observation of cognitive functioning included aspects such as attention and understanding. Possibly, feelings of anxiety related to the recent experience of the disaster may produce a heightened level of arousal and alertness, which in turn leads to improved cognitive functioning. The finding that after this brief improvement cognitive functioning declined at a normal rate, shows correspondence with what is known about the effects of stress: in the short term, stress may have beneficial effects on functioning, whereas continued stress has detrimental long-term effects.25
As the LASA study was not initiated to address the specific issue dealt with in this study, there are a number of limitations. First, the available exposure measure, residential distance from the disaster, is used as a proxy for actual experience of the disaster. This approach assumes that the event of the crash affects all respondents in a distance-based subsample equally, and ignores differences in subjective experience and coping. Nevertheless, such differences are highly likely, considering that some actually suffered the loss of a loved one or of their home, whereas others only heard from the suffering of distant neighbours. The lack of more detailed exposure measures in our study may result in an underestimation of the effects. Second, the subsample living nearest to the site of the disaster was small. Moreover, it suffered from considerable attrition, so that for the examination of longer-term effects only 21 subjects were left in the near Bijlmer subsample. Although mortality rates did not differ among the subsamples, non-mortality attrition did show differences and may be a source of bias. However, both mortality and non-mortality attrition were less influenced by health status in the Bijlmer residents than in the subjects living outside the Bijlmer suburb. This implies that the continuing participants in the Bijlmer subsamples were not excessively selective with respect to the outcome variables. Third, this study was necessarily based on the general health indicators that were available both at the pre-disaster and post-disaster data collection cycles. Thus, specific aspects of health that may be more closely related to the experience of a disaster, such as post-traumatic stress disorder, could not be investigated. On the other hand, the availability of general health indicators may be considered as a strength of this study, as there is very little research on changes in general health in relation to a disaster.
In all, the limitations of this study are offset by the availability of a pre-disaster measurement. Inevitably, studies that are explicitly initiated to investigate the health effects of a disaster have post-disaster measurements only, which hampers interpretation of differences between subjects exposed and not exposed.
A final note concerns the older age of the sample studied. The question may be raised whether results among older residents can be applied to the total Bijlmer population. First, the population aged 55 and over forms a relatively small minority in the Bijlmer suburb. Second, older persons may show different reactions to a disaster than younger ones.26 On the one hand, older persons are more frail and may thus be more vulnerable to the effects of a disaster than younger persons.27 On the other hand, older persons may be less reactive to stress events (‘maturation’), or, in their longer lives, may have had to deal with similar disasters, which may make it easier to cope with the more recent one (‘inoculation’).28 A study among both younger and older adults, could clarify age differences in these possible reactions.
In conclusion, this study shows that older persons are likely to experience negative health changes in the wake of a disaster, over and above the negative health changes that occur normally with aging. Because a disaster affects all people living in an affected area, its public health consequences may be considerable. Nevertheless, it should be noted that the disaster-related health declines were small, and most obvious in the ability to perform actions (such as mobility), but were neither observed in disability in daily functioning, nor in self-perceptions of health. These findings may be interpreted as indicative of substantial resilience in older persons, despite their additional health problems.
Key points
|
|
Acknowledgments |
|---|
The Longitudinal Aging Study Amsterdam is largely funded by the Netherlands Ministry of Health, Welfare and Sports. This study was also supported by the Medical Research Aeroplane disaster Bijlmermeer (MOVB).
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
1 Katz CL, Pellegrino L, Pandya A, Ng A, DeLisi LE. Research on psychiatric outcomes and interventions subsequent to disasters: a review of the literature. Psychiatry Res 2002;110:201–17.[CrossRef][ISI][Medline]
2 McFarlane AC. The longitudinal course of posttraumatic morbidity, the range of outcomes and their predictors. J Nervous Mental Dis 1988;176:30–9.[ISI][Medline]
3 McFarlane AC, Papay P. Multiple diagnosis in posttraumatic stress disorder in the victims of a natural disaster. J Nervous Mental Dis 1992;180:498–504.[ISI][Medline]
4 Escobar JL, Canino G, Rubio-Stipec M, Bravo M. Somatic symptoms after a natural disaster: a prospective study. Am J Psychiatry 1992;149:965–7.
5 Shalev A, Bleich A, Ursano RJ. Posttraumatic Stress disorder: somatic comorbidity and effort tolerance. Psychosomatics 1990;31:197–203.
6 Engel CC Jr. Outbreaks of medically unexplained physical symptoms after military action, terrorist threat, or technological disaster. Mil Med 2001;166:47–8.[Medline]
7 Barsky AJ, Borus JF. Functional somatic syndromes. Ann Intern Med 1999;130:910–21.
8 Wessely S. Ten years on: what do we know about the gulf war syndrome? King's College Gulf War Research Unit. Clin Med 2001;1:28–37.[ISI][Medline]
9 Nimnuan C, Rabe-Hesketh S, Wessely S, Hotopf M. How many functional somatic syndromes? J Psychosom Res 2001;51:549–57.[CrossRef][ISI][Medline]
10 Feder A, Olfson M, Gameroff M, et al. Medically unexplained symptoms in an urban general medicine practice. Psychosomatics 2001;42:261–8.
11 Nimnuan CH, Hotopf M, Wessely S. Medically unexplained symptoms. An epidemiological study in seven specialties. J Psychosom Res 2001;51:361–7.[CrossRef][ISI][Medline]
12 McFarlane AC, Atchison M, Rafalowicz E, Papay P. Physical symptoms in post-traumatic stress disorder. J Psychosom Res 1994;38:715–26.[CrossRef][ISI][Medline]
13 Friedman MJ, Yehuda R. Post-traumatic stress disorder and comorbidity. In: Friedman MJ, Charney DS, Deutsch AY, Editors. Neurobiological and clinical consequences of stress: from normal adaptation to PTSD. Philadelphia: Lippencott-Raven Publishers, 1995:429–45.
14 Reijneveld SA. The impact of the Amsterdam aircraft disaster on reported annoyance by aircraft noise and on psychiatric disorders. Int J Epidemiol 1994;23:333–40.
15 Knipscheer CPM, De Jong Gierveld J, Van Tilburg TG, Dykstra PA, Editors. Living arrangements and social networks of older adults. Amsterdam: VU University Press, 1995.
16 Deeg DJH, Knipscheer CPM, Van Tilburg W, Editors. Autonomy and well-being in the aging population: concepts and design of the Longitudinal Aging Study Amsterdam. NIG Trend studies 7. Bunnik: Netherlands Institute for Gerontology, 1993.
17 Deeg DJH, Van Tilburg T, Smit JH, De Leeuw ED. Attrition in the Longitudinal Aging Study Amsterdam: the effect of differential inclusion in side studies. J Clin Epidemiol 2002;55:319–28.[CrossRef][ISI][Medline]
18 Van Sonsbeek JLA. Methodological and substantial aspects of the OECD indicator of chronic functional limitations. Maandbericht Gezondheid (NCBS) 1988;7/6:4–17.
19 Kriegsman DMW, Deeg DJH, Van Eijk JTM, Penninx BWJH, Boeke AJP. Do disease specific characteristics add to the explanation of mobility limitations in patients with different chronic diseases? A study in the Netherlands. J Epidemiol Community Health 1997;51:676–85.[Abstract]
20 Anderson JS, Sullivan F, Usherwood TP. The Medical Outcomes Study instrument (MOSI)—use of a new health status measure in Britain. Fam Pract 1990;7:205–18.
21 Van Sonsbeek JLA. The self-rating of health: methodological effects of the rating of health in health interview surveys. Maandbericht Gezondheid (NCBS) 1991;10(9):15–23, (in Dutch).
22 Verbrugge LM, Jette A. The disablement process. Soc Sci Med 1994;38:1–14.[CrossRef][ISI][Medline]
23 Linn BS, Linn MW. Objective and self-assessed health in the old and very old. Soc Sci Med 14A 1980:311–4.
24 Borawski EA, Kinney JM, Kahana E. The meaning of older adults' health appraisals: congruence with health status and determinant of mortality. J Gerontol Soc Sci 51B 1996:157–70.
25 Kraaij V, Arensman E, Spinhoven P. Negative life events and depression in elderly persons: a meta-analysis. J Gerontol Psychol Sci 57B 2002:P87–P94.
26 Norris FH, Kaniasty K, Conrad ML, Inman GL, Murphy AD. Placing age differences in cultural context: a comparison of the effects of age on PTSD after disasters in the United States, Mexico, and Poland. J Clin Geropsychol 2002;8:153–73.[CrossRef]
27 Havens B, Hall M. What did we learn with pre and post flood data with those age 75 and older? Gerontologist 1999;39(Special Issue I):340.
28 Knight BG, Gatz M, Rose T, Bengtson V. Age and emotional response to the Northridge earthquake: a longitudinal analysis. Gerontologist 1999;39(Special Issue I):340.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||