Psychological health and coping strategy among survivors in the year following the 2008 Wenchuan earthquake

Authors


Jiuping Xu, PhD, Uncertainty Decision-making Laboratory, Business School, Sichuan University, No. 29 Jiuyanqiao Wangjiang Road, Chengdu 610064, China. Email: xujiuping@scu.edu.cn

Abstract

Aim:  The powerful earthquake of 12 May 2008 wrought incalculable havoc on lives and properties in Wenchuan, Sichuan Province, China. The catastrophic earthquake not only created tremendous changes in the external environment, but also caused stress and difficulties for the people in the affected areas which were felt long after the event. In this study, we attempt to clarify the correlation between coping strategies and psychological well-being among survivors across sex and levels of exposure.

Method:  A total of 2080 survivors from 19 counties freely participated in the survey which used self-report psychological questionnaires, the Short Form-12, version 2 Scale and Coping Scales. We estimated regression models to identify the coping factors associated with the presence of mental symptoms after the disaster.

Results:  Four main factors (middle-age, low educational level, low monthly income, and high exposure) were significantly related to poor health. Highly exposed survivors tended to problem-avoidance, fantasy, self-blame and seeking assistance, which was significantly different to those lowly exposed. Women tended to be more vulnerable than men and exhibited problem-avoidance and self-blame. Six coping styles were significant determinants and predicted 64.2% of health.

Conclusion:  Post-disaster mental health recovery intervention, including early identification, ongoing monitoring, sustained psychosocial support and more mental health services, are required for the high-risk population, especially for women.

IN MAY 2008, an extra-large earthquake occurred in Wenchuan, China, spreading to Sichuan, Gansu, Shaanxi, Chongqing, Yunnan and 10 other provinces (including districts and municipalities), with a total affected area of about 500 000 km2. The earthquake epicenter was located at latitude 31.021°N and longitude 103.367°E. The magnitude reached 8.0 on the Richter scale with a maximum intensity of 11 degrees. The earthquake caused 69 227 deaths, including 68 636 in Sichuan province; there were 374 643 injuries, and 17 923 people were missing. The earthquake left about 4.8 million people homeless. Approximately 15 million people lived in the affected area. The catastrophic earthquake caused people in the affected areas traumatic stress disorders for a long time.

When an unexpected or an emergency event happens, people's behavior in the face of natural hazards is deeply influenced by cultural, social, economic, and political contexts.1 The body of literature on post-trauma psychology has focused on survivors in economically developed countries over the past 20 years.2 Pynoos et al. found that 70% of the survivors showed post-traumatic stress symptoms 1.5 years after the 1988 Armenian Earthquake.3 Post-traumatic stress disorder (PTSD) is probably the most frequently discussed topic in disaster psychology theory.4–6 The prevalence of PTSD reported in victims of earthquake trauma ranges from 10% to 87%.7 The results in the above papers showed that severe earthquakes can cause long-lasting morbidity.8,9 Reviews of disaster studies have concluded that large-scale community traumas can result in a significant increase in psychological problems in the short-term and can have significant negative physical and mental health consequences for years post-disaster.9–11

Some researchers have contended that not all adults exposed to disasters develop PTSD, and some persons recover quickly from these experiences.12,13 However, some factors appeared to associate with increased risk of psychological problems,2,11 such as trait anxiety and negative affect.10,14 A number of factors have been identified to be predictive of development of psychological problems, such as sex,15,16 accumulation of multiple stressors,10,17 severity of trauma exposure,14,18,19 and coping.17,19 Coping is described as an individual's efforts to master demands (conditions of harm, threat or challenge) that are appraised (or perceived) as exceeding or taxing his or her resources.13 Thus it is likely that one's coping strategy will be intimately associated with the severity of distress one experiences. For example, effective use of coping strategies may protect a person from cognitive, environmental, and biologic factors that may bring about symptoms of anxiety. Retrospective research showed that coping strategies are related to PTSD.16,17,20,21 In addition, the occurrence of a natural disaster in a community is a large-scale environmental stressor.

Because of the altered environment, people have to adapt quickly to new circumstances, often facing loss of friends, family, and possessions.9,11,22 Furthermore, when an earthquake occurs, the victims have to live with the fear of potential recurrence, or aftershocks, as several earthquakes often occur in succession. This sequence of events may even affect an individual's ability to regulate, identify, and express emotions, as well as affect later development,23 and has a negative effect on the individual's core identity and ability to relate to others.24 On the basis of psychosocial stress theory,25,26 some aspects of survivors' lives can strengthen or weaken their ability to cope with a community disaster. More specifically, demographic characteristics, such as socioeconomic status or sex, and a deferent coping strategy can add to or reduce the distress levels of individuals undergoing a traumatic event.2,25,26

Most previous disaster studies are not based on random or probability samples of well-defined populations.2,11 Indeed, many are restricted to persons seeking medical, psychological or personal assistance. Additionally, most sample sizes are small: for example, the median size of the 169 samples included in the Norris reviews2 is 149. The strength of this study is a large random sample representative of adult survivors in hard-hit regions and the usage of standardized scale of coping.

The focus of the study was to examine the mental health state and coping strategy of survivors 1 year after the disaster based on the context of Chinese society and culture. We use data from a survey of 19 counties in the disaster area which provided a unique opportunity to identify whether a broad range of risk factors are associated with vulnerability or resilience to depressive symptoms in the aftermath of this catastrophic event.

METHODS

Data collection

The data for the present study came from investigations into the consequences of the Wenchuan earthquake. According to the stricken counties list published by the China Earthquake Administration, 19 heavily stricken counties surrounding the epicenter in two provinces (Sichuan and Shaanxi) were selected randomly. In Sichuan Province, the 18 counties were Dujiangyan, Pengzhou, Chongzhou, Shifang, Mianzhu, Jiangyou, Anxian, Pingwu, Beichuan, Jiange, Qingchuan, Hanyuan, Wenchuan, Lixian, Maoxian, Songpan, Heishui, and Xiaojin. In Shanxi province, Lueyang was selected.

From July to September 2009, master's level psychology students working as research assistants approached the participants in their own homes or in temporary accommodation. Houses and tents were randomly selected on the basis of the total number. Participants were given information orally about the study's purpose and oral consents were obtained before each interview. One respondent within each selected house or tent was randomly selected according to birth date. Those who experienced the violent earthquake and were aged from 18 to 65 years were available to be interviewed. Some individuals declined the interview because they were wary of such earthquake surveys and avoided talking about the event. If this was the case, the next closest one was invited instead. Individuals with mental retardation and major psychoses (e.g. schizophrenia, major depressive disorder, organic mental disorders) were excluded. Each item was read to those participants who had difficulty understanding the printed questionnaires, or who had a low level of education. A total of 2300 individuals were involved in this survey and 2080 completed the questionnaire, with a response rate of 90.4%.

The study was designed in accordance with the tenets of the Declaration of Helsinki and was approved by the ethics committee of Sichuan University. The investigation complies with the principle of voluntariness.

Measures

The sociodemographic characteristics of the study included five variables: age, sex, ethnic group, education level, and income level. Age was coded to the nearest year. Sex and ethnic group were dummy variables. Education was coded into four categories, including doctor, master, bachelor, and others. Income was coded into four categories, including under 1000 RMB, 1000–2000 RMB, 2000–3000 RMB, and more than 3000 RMB per month.

Psychological health was evaluated using the Short Form-12, version 2 (SF-12-v2) (Cronbach's alpha = 0.85). SF-12-v2 is a suitable measure for large group studies (greater than n = 500) where information on the Short Form-36 Health Survey Summary Scores (Physical Component Summary + Mental Component Summary) is required, and it was proved valid for the Chinese in previous studies.27,28 High item scores reflect better health. A T-scores algorithm was designed to convert psychological items to standard scores so that the scale would have scores with a mean close to 50, a standard deviation close to 10, and be uncorrelated with each other. Although both scales contain all 12 items, the physical health measure (SF-12-v2 physical component, range 17.15–77.63) emphasized physical functioning, role functioning, body pain, and general health status over the previous 30 days. The psychological health measure (SF-12-v2 mental component, range 17.04–77.19) emphasized vitality, social functioning, emotional functioning, and mental health status over the previous 30 days.29 The SF-12-v2 scale has good reliability and validity, correlates well with clinical assessments of physical and mental health,29–31 and has been used in numerous studies worldwide, including China.27,28,30

Assessment of exposure was the sum of multiple choices to measure whether participants had encountered the following situations as a result of the earthquake: personal injuries; death or injury of family members, relatives or friends; loss of or damage to personal or family property; witnessing other people who were seriously injured or killed; changing jobs after the earthquake; and relocation to temporary shelters. The higher scores represented greater earthquake-related exposure. Based on an examination of the frequency distribution, we coded respondents into three exposure degrees: low exposure (0–1 event), moderate exposure (2–3 events), and high exposure (more than 4 events).

The coping strategies assessment was evaluated using the Coping Scale (CS), which is intended as a widely applicable self-report measure for situational coping and encompasses the avoidance, problem-solving and seeking social support strategies. 32 The CS consists of three subscales: Mature Coping (including two factors: one factor was asking for help from others; another was problem-solving), Immature Coping (including three factors: self-blame, problem-avoidance and fantasy), and Mixed Coping (only one factor: rationalization). Each subscale comprised 1–3 factors. Each factor had 3–5 items. Each item could be answered with a score on a 5-point Likert scale where 1 = not sure at all and 5 = very sure. A compound score for each factor was estimated by averaging the scores of all items in the factor. A grand score for each subscale was obtained by averaging all the compound scores in the subscale. CS grand score produced a Cronbach's alpha of 0.879 and Split-Half of 0.865 in a reliability test.

Statistical analysis

Descriptive statistics, such as mean, SD, and ranges for the variables were calculated and presented. One-way anova analyses were performed to compare the mental health scores of different groups of participants. Hierarchical multiple regression analyses were performed to predict physical and mental health outcomes, respectively, from six coping style variables. The regression analysis proceeded in six steps to assess the explanatory power of each variable. These results revealed the unique effects of the independent variables, controlling for other variables in the model. All statistical procedures were completed using spss 16.0 (spss, Chicago, IL, USA).

RESULTS

The demographic characteristics of the study sample are given in Table 1. The mean age of the 2080 participants at the time of the interview was 38.24 ± 8.82 years (ranging from 18 to 65 years). The majority (i.e. 59%) of the subjects were male. Overall, 53% of the respondents had a relatively low education level. Apart from the majority ethnic group (Han, 80.5%), the following ethnic groups participated in our survey: Tibetan (7.1%), Qiang (10.1%), Hui (1.8%), and others (like Tujia and Yi etc., 0.5%). In terms of personal income, 83.9% earned less than 2000 RMB per month, about 16.1% earned more than 2000 RMB per month.

Table 1.  Sociodemographic characteristics of the study sample (n = 2080)
 n%
Sex  
 Male122759.0
 Female85341.0
Age groups (years)  
 18–24743.5
 25–3458428.1
 35–4485241.0
 45–5449123.6
 55–65793.8
Monthly income  
 <1000 RMB38518.5
 1000–2000 RMB136065.4
 2000–3000 RMB26412.8
 >3000 RMB713.3
Ethnic group  
 Han167480.5
 Tibetan1477.1
 Qiang21110.1
 Hui371.8
 Others110.5
Education degree  
 Doctor120.6
 Master351.7
 Bachelor93845.1
 No degree109552.6

Grouped by demographic variables, the scores for physical and mental health status are shown in Table 2. The anova was used to test the hypothesis that several means in the same group were equal, the two-tailed χ2P-values were given out for each group, which are also shown in Table 2. Note that a higher mean score showed a better degree of well-being. The difference among age subgroups was significant. The subgroup ranging from 35 to 44 reported the lowest score in health. Respondents younger than 34 reported higher scores in physical health than mental health, meanwhile respondents older than 34 reported higher scores in mental health than in physical health. When considering sex, only mental health was statistically significant where men reported higher than women, but there was no significant difference between men and women in physical health and overall health. Among ethnic subgroups, there was a distinct difference in physical health and in overall health, but no difference in mental health. The Hui ethnic group had the highest overall health scores of all ethnic groups. Except for the Tibetans, all ethnic subgroups had higher scores in physical health than mental health. Educational level groups were statistically significant in three aspects of well-being. Among those subgroups, respondents with a PhD degree had the highest physical and mental health scores compared with other lower education degree subgroups. Respondents with a master degree reported the second highest score on this item. Respondents with a bachelor degree had a small difference in the health score compared with respondents with lower degree level of education. Survivor's resilience post-disaster varied significantly with income, with those on a higher economic level coping better.

Table 2.  Scores of physical and mental health status on each sociodemographic variable (n = 2080)
VariablesSF-12 physical healthSF-12 mental healthSF-12 health
Mean (SD)P-valueMean (SD)P-valueMean (SD)P-value
  1. *P < 0.05; **P < 0.01; ***P < 0.001.

Age (years)      
 18–2451.73 (8.62)**49.80 (9.41)*102.11 (15.87)*
 25–3451.05 (10.62)50.56 (10.41)101.60 (19.11)
 35–4448.99 (10.12)49.06 (9.80)98.38 (18.33)
 45–5450.33 (8.92)50.81 (9.62)101.70 (16.11)
 55–6549.12 (10.30)51.59 (11.21)100.79 (20.47)
Sex      
 Male50.10 (10.03)0.51150.89 (10.04)***100.88 (18.20)0.075
 Female49.77 (9.73)48.69 (9.60)99.17 (17.47)
Ethnic group      
 Han49.86 (10.09)**49.78 (10.11)0.16199.86 (18.25)*
 Tibetan48.59 (8.71)50.97 (9.83)100.16 (16.40)
 Qiang50.25 (9.59)49.36 (9.23)99.71 (17.46)
 Hui56.04 (6.45)53.58 (7.03)100.73 (12.10)
Education level      
 Doctor70.20 (10.59)***65.92 (12.50)***131.85 (25.06)***
 Master54.44 (8.23)53.26 (7.90)106.65 (14.47)
 Bachelor50.17 (10.61)49.65 (10.41)99.97 (19.78)
 No degree49.42 (9.18)50.06 (9.53)99.99 (16.25)
Monthly income      
 <1000 RMB48.71 (10.60)***48.96 (9.21)***97.53 (17.47)***
 1000–2000 RMB49.67 (10.09)49.38 (10.32)99.30 (18.65)
 2000–3000 RMB52.79 (8.73)52.96 (8.39)106.32 (15.35)
 >3000 RMB53.05 (7.50)55.58 (10.16)109.18 (16.68)

Table 3 shows the means, standard deviations and the F-test scores for different health and coping strategies exposure degrees. There were significant statistical differences in the three degrees on health and coping strategy scores. The differences in health compound scores demonstrated the differences in the mental states of the three exposure categories. Participants experiencing less exposure events had significantly higher scores than those experiencing more exposure events on the entire health compound scores (P < 0.001). The differences in coping strategy compound scores demonstrated differences in coping styles among the three exposure categories. The results showed that there were no differences in rationalization and problem-solving coping styles (P > 0.05), with participants who experienced high exposure having significantly higher scores than those experiencing low exposure on problem avoidance (2.43 vs 2.27), fantasy (3.57 vs 3.20), self-blame (2.54 vs 2.10), and asking for help from others (3.27 vs 3.03) coping scores (P < 0.05).

Table 3.  Health and coping strategy in differences of exposure (n = 2080)
VariablesHigh exposure (n = 392)Moderate exposure (n = 766)Low exposure (n = 922)F(df1,df2)P-value
MeanSDMeanSDMeanSD
  1. *P < 0.05; ***P < 0.001.

Health        
 Physical Health47.879.8149.4710.0250.6110.037.21(2,2077)***
 Mental Health47.269.9049.209.6150.6910.338.72(2,2077)***
Coping strategy        
 Avoiding problem2.430.942.400.882.270.814.12(2,2077)*
 Fantasy3.570.863.520.773.200.643.45(2,2077)*
 Self-blaming2.540.912.540.922.100.823.85(2,2077)*
 Ask for help from others3.271.083.090.973.030.913.62(2,2077)*
 Rationalization2.840.922.810.782.790.881.40(2,2077)0.26
 Solving problem3.360.863.370.853.310.820.31(2,2077)0.73

Table 4 shows the means, SD and the t-test scores for sex differences on health, exposure and coping strategy. There were no statistically significant sex differences on exposure. For the compound health scores, the results showed that, consistent with Table 1, men had higher scores than women as regards mental health (50.89 ± 10.02 vs 48.69 ± 9.58, t(2078) = 4.57, P < 0.001). For the coping strategy compound scores, the results showed that women had higher scores than men on problem-avoidance (2.54 ± 0.84 vs 2.37 ± 0.92, t(2078) = 3.98, P < 0.001) and self-blame coping styles (2.62 ± 0.90 vs 2.49 ± 0.89, t(2078) = 2.95, P < 0.005), while men had higher scores than women on fantasy (3.63 ± 0.82 vs 3.48 ± 0.81, t(2078) = 3.81, P < 0.001) and problem-solving coping styles (3.42 ± 0.85 vs 3.29 ± 0.86, t(2078) = 3.08, P < 0.005).

Table 4.  Health, exposure and coping strategy in sex differences (n = 2080)
VariablesMale (n = 1219)Female (n = 861)t(df)P-value
MeanSDMeanSD
  1. **P < 0.01; ***P < 0.001.

Health      
 Physical Health50.1010.0249.769.720.65(2078)0.511
 Mental Health50.8910.0248.699.584.57(2078)***
Exposure1.740.751.780.76−1.15(2078)0.249
Coping strategy      
 Avoiding problem2.370.922.540.84−3.98(2078)***
 Fantasy3.630.823.480.813.81(2078)***
 Self-blaming2.490.892.620.90−2.95(2078)**
 Ask for help from others3.261.053.161.091.92(2078)0.055
 Rationalization2.850.852.770.851.65(2078)0.099
 Solving problem3.420.853.290.863.08(2078)**

A hierarchical regression model was used to assess potential predictors of health. A block of sociodemographic characteristics, exposure degree and coping strategy compound scores were entered as independent variables, with the health grand score as a dependent variable. Table 5 presents the contribution of each of the independent variables towards predicting the severity of mental problems. It displays the adjusted R2 and the standardized regression coefficient (Beta). The predictor variables explained 64.2% of the total variance in the mental health grand score (adjusted R2 = 0.642, F[6, 2073] = 379.669, P < 0.001) in six steps. The mental health grand scores are high when the following are found: the use of fantasy, help, rationalization, and problem-solving coping styles, and the mental health grand scores are low with the use of avoidance and self-blame coping styles. This means that the use of avoidance and self-blame coping styles can predict more severe mental problems.

Table 5.  Six-step hierarchical regression analysis on coping strategy to predict mental health (n = 2080)
PredictorAdjusted R2BetatdfP-value
  • ***

    P < 0.001.

Step 1: Avoiding problem0.331−0.575−25.031***
Step 2: Fantasy0.5260.48622.852***
Step 3: Self-blaming0.571−0.223−11.573***
Step 4: Ask for help0.5970.1789.014***
Step 5: Rationalization0.6290.19910.545***
Step 6: Solving problem0.6420.1536.876***

DISCUSSION

Earthquake disasters can result in enormous human and economic costs.2 At the community level, they may destroy resources, damage infrastructure, and place overwhelming demands on local authorities.2,10 On an individual level, they may have deleterious consequences for psychological health.5–8 The findings of this study 1 year after the Wenchuan earthquake highlight the survivors' psychological symptoms. Our results on the psychology and coping styles after the Wenchuan earthquake are comparable with post-disaster rates reported elsewhere.7,16,17,19,21 Many associated negative factors have been identified in the cross-sectional studies:1,2 high exposure,18,19 female sex,15,16 middle-aged,33 low income,2,26 and low educational level.23,26 Although many of the findings are consistent with previous studies,2,6,9 some varied because of the earthquake intensity, the types of exposure, the sampling selection, the time elapsed since the earthquake, and the measures used.

Many studies have documented the correlation between age and mental health. Some studies reported that older people have higher recovery resilience,33–35 while others report the contrary.33,36,37 Our study observed that the middle-aged (aged 35–44) in the aftermath of a disaster are a risk factor, which is consistent with previous research.36,38 One possible reason is the burden perspective hypotheses,2,36 which suggests that middle-aged adults experience poorer coping capacity than others because their responsibilities to society (e.g. working) and to the family (e.g. often providing support to both children and parents) render them more psychologically vulnerable in the aftermath of the disaster.22,23 Another report from the Wenchuan earthquake suggests that survivors aged 41–50 were more severely affected by direct exposure to the effects of the earthquake and gives similar reasons.38

Not every study looked for sex effects, and not every study that looked for them found them. Outcomes of some studies were that women and girls were twice as likely to develop PTSD as men and boys.15,16,39,40 Our study reported that women had greater emotional distress, trauma, and mental health problems than men, but there was no significant difference in physical health.41 One possibility is the difference of sex on a subjective measure of seismic exposure,42 so women might experience greater levels of fear and anxiety regarding their own safety and that of their family members. Another possibility is higher emotional stress in women before the earthquake. However, few explanations have been empirically tested, making it difficult to draw any specific conclusions concerning sex differences.

Some studies found numerous associations between ethnicity and mental health.2,43–45 In our study, we found the Qiang ethnic group reported the lowest total health score. Our findings showed that minority groups most often had poorer health than persons of majority group status. This may result from the fact that the Qiang ethnicity was located in one of the most severely affected counties and the severity of exposure accounted for much of the Qiang minority group members' higher post-traumatic stress.45,46 Another possibility is their weak mental health before the earthquake.

The influence of socioeconomic status was assessed in our study. One year after the disaster, participants with the highest level of education or highest income reported high scores for physical and mental health (Doctor: physical score 70.20, mental score 65.92; Income over 3000 RMB: physical score 53.05, mental score 55.58), which indicated that higher socioeconomic status had an especially positive effect on physical health. Previous studies proved that a lower socioeconomic status was consistently associated with greater post-disaster distress.2,4,47–49 As Asgary and Willis noted, a higher level of education may cushion the impact of the disaster and enhance recovery by access to better information and relevant resources.50

Another purpose in our research was to assess the correlation between individual well-being and coping strategies after the exposure impact of the Wenchuan earthquake, and we found that highly exposed participants had a worse mental state than those with low exposure. The highly exposed participants tended more to use problem-avoidance, fantasy, self-blame, and asking for help from others as coping styles which may be related to the disaster exposure degree. It is known that the degree of disaster exposure,18,30 bereavement,4,42 injury to self or family members,17,48 property damage or financial loss51 and relocation52 increase the vulnerability to mental symptoms. According to the investigation, the basic information indicated significantly that the highly exposed participants experienced the following: personal injuries; the death or injury of family members, relatives or friends; the loss of or damage to personal or family property; being a witness to other people seriously injured or killed; and being relocated to temporary shelters. Those victims suffered a prolonged period of helplessness with little contact with the outside world, panic and physical hardship, which may be the reason for the more severe mental problems and negative coping styles.

Although individuals influence their psychological outcomes for the better according to their ways of coping,19,53 sex differences of the coping strategy were statistically significant in the study. More women adopted avoidance and self-blame coping styles, while men adopted problem-solving and fantasy coping styles. Therefore, we could speculate that the men who were apt to adopt the problem-solving and fantasy coping styles may have fewer mental problems. This may also be related to sex characteristic differences,15,16 or may be eliminated by a higher trauma exposure.18,19,54 However, previous studies varied on sex coping. Eschenbeck et al. indicated that in stressful situations, women were likely to seek social support and problem-solving while men tended to avoidance.55 In Matud's study, women were found to have more emotional and avoidance coping styles than men, but less rational and detachment coping. Further study on sex coping is needed.16

With hierarchical regression analysis, avoidance, fantasy, self-blame, asking for help, rationalization and problem-solving coping styles were taken as a predictor to identify the mental health outcome 1 year after the earthquake. Our study suggests that fantasy, asking for help, rationalization and problem-solving had a positive effect on mental symptoms, while problem-avoidance and self-blame had negative effects. The avoidance coping Beta value (−0.575) was problematic, consistent with previous studies.19,53 It could be more important to help the survivors to reduce their avoidance, which seems to play the most important role in decreasing mental trauma symptoms after an earthquake. The use of avoidance and self-blame may enhance vulnerability to mental problems.56,57 Our study suggests that fantasy has the most positive correlation to mental health, consistent with some personality research suggestions that fantasy and daydreams were positive for life enhancement.58 Thus, specific intervention and attention that could reduce mental symptoms may be helpful. Because survivors rarely seek psychiatric help, many special measures need to be designed, such as home visits and mobile clinics.59 Mental health service providers should work to minimize survivors' victimization experiences, improve their regulatory abilities, and promote alternatives to negative coping in victims following disasters.60 Mental health recovery intervention like early identification, ongoing monitoring, sustained psychosocial support and mental health services are required for the high-risk population. This study will be useful for psychologists to provide effective intervention and prevention to post-traumatic mental problems in the future.

Similar to other studies, our results need to be viewed with regard to the limitations. First, the study design was a cross-sectional study, which did not provide valid information about the previous mental health of the subjects. There was no measure of pre-disaster resilience and vulnerability factors for comparison to post-disaster. Second, selection bias may exist in the data because of the sample and recruitment method. Third, a self-report instrument was used, so the participants may be over- or underreported. Self-reporting is subject to recall bias, and some participants might have attributed psychological symptoms to the earthquake that were actually more associated with other traumatic events, which may limit the strength of the findings.

ACKNOWLEDGMENTS

We would like to thank all of the interviewees who showed great patience in answering the questionnaires. This research was supported by Major Bidding Program of National Social Science Foundation of China (Grant no. 08&ZD009) and also partially sponsored by the project of Investigation Propaganda Department of China Association for Science and Technology (Grant no. 2009DCYJ12).

Ancillary