There is compelling evidence suggesting that exposure to a massive earthquake increases the rates of posttraumatic stress disorder (PTSD) among surviving children and adolescents, with a prevalence of PTSD reported to vary between 12% and 70% within 2 years.Reference Blanc, Bui and Mouchenik 1 Studies have shown a decline in the severity of PTSD symptoms among adolescents as time passes after the earthquake; however, a few studies have identified intractable symptoms that develop into chronic PTSD.Reference Goenjian, Walling and Steinberg 2 Moreover, teens experience dramatic changes in physiological and psychological functions. As their psychological development lags behind their physical development, adolescents often experience many internal contradictions that increase their susceptibility to common mental health problems (CMHPs), such as anxiety and depression.Reference Tang, Guo and Huang 3 It is unclear whether CMHPs among adolescents aggravate or hinder improvement in PTSD symptoms.
Some previous researchers have documented that earthquake-related factors, such as home damage, loss of possessions, and loss of family members, aggravate the symptoms of PTSD among adolescents.Reference Alisic, Zalta and Wesel 4 These risk factors of PTSD are largely reported within 2 years after disasters among adolescents whose homes have not yet been rebuilt. Some psychological factors, such as self-esteem, psychological resilience, social support, and positive coping, have been reported to protect against PTSD.Reference Mann, Hosman, Schaalma and de Vries 5 However, a few studies have found that these protective psychological factors for PTSD may change when adolescents grow up.
The devastating earthquake measuring 8.0 on the Richter scale that struck Wenchuan in 2008 in Sichuan Province, China (http://www.512gov.cn/GB/123057/8074265.html), caused more serious destruction and mental harm than other disasters. Therefore, it is necessary to determine the long-term impact of this earthquake in adolescents to implement effective countermeasures for improving their mental health. Some studies have documented a prevalence of PTSD of 15.8% to 60.1% among adolescents within 2 years after the Wenchuan earthquake.Reference Fan, Zhang and Yang 6 However, no studies on the prevalence of PTSD and its correlates among adolescents surviving this earthquake have focused on a period longer than 4 years.
Two objectives of this study were (1) to identify the prevalence of PTSD at 53 months after the earthquake and determine the contributions of demographic variables and earthquake experiences to the risk of PTSD, and (2) to assess changes in the psychological determinants influencing PTSD among the surviving adolescents.
METHODS
Study Participants
Adolescents were recruited from 2 junior high schools using stratified cluster sampling. One school was in the most earthquake-damaged area of Mianzhu County, located in the seismic belt approximately 50 linear miles from the epicenter. The other school was in the Longquanyi district of Chengdu City, an area that experienced only slight damage from the earthquake, located outside of the seismic belt approximately 70 linear miles from the epicenter. The school in Mianzhu County is a combination of 3 destroyed schools. All of the buildings in this school were rebuilt at a new site after the earthquake. The school in Longquanyi district, Chengdu City, experienced damage to a small number of buildings and equipment, but they were usable after repair. All participants had personally experienced shaking buildings and broken roads. Some from the school in Mianzhu had escaped from collapsed buildings, and a few had even encountered people who were injured or died.
The first baseline survey included 1,256 participants with a mean age of 12.7 years (range=10.9~15.4 years) who were in the first semester of grade 7 of junior high school in October 2010, that is, 29 months after the earthquake. Follow-up surveys were then conducted among the same participants every 6 months until October 2012, 53 months after the earthquake, when the participants were in the first semester of grade 9. The PTSD survey was administered at the fifth follow-up in October 2012 only. In this study, we extracted participants’ demographic characteristics, earthquake experiences, and psychological variables from the baseline survey (grade 7) and psychological variables from the third survey (grade 8, October 2011, sample size of 1,227). We then linked these data with PTSD measures and the corresponding correlates collected in the fifth survey (grade 9, sample size of 1,245). The 1,125 individuals who participated in surveys 1, 3, and 5 were included in this study.
Field Iinvestigation
Psychological counsellors were involved in the field investigation to prevent any emergencies from occurring during the survey. Moreover, 2 pilot surveys were performed, and the survey time was controlled. At each survey, all of the participants simultaneously completed the questionnaire by themselves in their own classrooms. All participants provided both oral and written consent before the study. Every investigator underwent thorough strict training. The study protocol was approved, and the whole process of the survey was supervised by the Quality Control Group and Medical Ethical Committee from Sichuan University, China (NO:ORG0004190/2010007).
Investigation Tools
Data were collected using self-report questionnaires that included demographics, earthquake exposure factors, and well-established psychological scales.
The 24-term PTSD Self-Rating Scale (PTSD-SS) is constructed based on the definition and diagnostic criteria of PTSD in the Manual of Mental Disorders, 4th Edition (DSM-IV), and the Chinese Classification and Diagnostic Criteria of Mental Disorders, 2nd Edition, Revised (CCMD-2-R).Reference Liu, Ma and Liu 7 Without clinical evaluation of participants in this study, we used a cutoff 50 points of a total PTSD-SS score to diagnose a PTSD-like mental disorder. In our study, Cronbach’s alpha coefficient was 0.97.
CMHPs were measured using the Mental Health Inventory of Middle-School Students (MMHI-60).Reference Wang 8 This instrument is a 60-item Likert scale including 10 subscales such as depression, anxiety, and paranoia. The total score is obtained by dividing the sum of the scores of the 60 items by 60 and is used to classify participants into 4 groups:<2=no CMHPs, 2~2.99=mild CMHPs, 3~3.99=moderate CMHPs, and ≥ 4=severe CMHPs. In this study, we aggregated the last 2 categories into 1 group termed moderate-to-severe CMHPs. Cronbach’s alpha coefficient in this study was 0.94.
The other scales included the Rosenberg Self-Esteem Scale (RSES) for self-esteem, Ego-Resiliency Scale (ER-89) for resilience, Social Support Rating Scale (SSRS) for social support, and Trait Coping Style Questionnaire (TCSQ:TCSQ-PC and TCSQ-NC) for coping style, with Cronbach’s alpha coefficients in this study as 0.81, 0.85, 0.71, 0.77, and 0.79, respectively.
Statistical Analysis
Frequencies, percentages, means, and standard deviations (SD) were calculated for the descriptive data analysis. Using no PTSD (PTSD score<50) as the reference, a univariate logistic regression model was used to calculate the odds ratios (ORs) and associated 95% confidence intervals (CIs) for correlations between PTSD and variables related to sociodemographic characteristics and earthquake experiences. T tests were performed to evaluate the differences in psychological covariates between the PTSD group and the no-PTSD group. A repeated variance analysis of each psychological covariate was conducted to compare differences between the 3 grades (grades 7, 8, and 9). Multicollinearity in the MMHI-60, RSES, ER-89, SSRS, TCSQ-PC, and TCSQ-NC scores was assessed before the multivariate analysis. Based on the previous analysis, we conducted a 3-block multivariate logistic regression using the Forward Conditional Stepwise method. Because of the lack of cutoff scores for the RSES, ER-89, SSRS, TCSQ-PC, and TCSQ-NC in the Chinese adolescent population, the scores on these scales were grouped into 3 categories based on their 33% and 67% percentiles to facilitate interpretation of their epidemiologic meaning. All tests were performed using SPSS Version 16.0 (IBM SPSS, Armonk, NY). The significance level was set at P ≤ 0.05 for 2-sided tests.
RESULTS
Of the 1,125 participants, 50.6% were girls. At the baseline survey, the mean participant age was 12.7 years (±0.5 SD). The mean PTSD score was 39.02 (±18.42 SD). The overall prevalence of PTSD was 23.4% (95% CI: 20.9%~26.0%) and that of moderate-to-severe PTSD (cutoff of 60 points) was 16.0% (95% CI: 13.9%~18.3%). In the no-PTSD group, the prevalence of moderate-to-severe CMHPs was 6.5%, 14.4%, and 8.8% in grades 7, 8, and 9, respectively, whereas the corresponding percentages in the PTSD group were 17.1%, 34.4%, and 58.0%, respectively.
The results of the univariate logistic regression model demonstrated that a higher prevalence of PTSD was associated with adolescents who were older, had poor physical health, experienced home damage or possession loss, had family members who died or were injured, and lived in the hardest-hit earthquake area with ORs of 1.42~1.85 (All P<0.05). Neither gender nor personal injury in the earthquake increased the prevalence of PTSD.
Table 1 demonstrates that compared with the no-PTSD group, adolescents with PTSD had less social support, lower ego-resilience, weaker positive coping, greater likelihood of negative coping, lower self-esteem, and worse CMHPs.
The Differences of Psychological Determinants Between the PTSD Group and No-PTSD Group
* In the same group of each variable, the same number of superscript signal means having no statistical significance. Different number of superscript means having statistical significance.
Table 2 indicates that older age and injury or death of family members were risk factors for PTSD. The presence of CMHPs was a strong risk factor for PTSD with ORs of 3.98 (95% CI: 2.34~6.78) to 17.67 (95% CI: 9.76~32.02). Self-esteem was consistently a protective factor for PTSD, regardless of increasing age. Positive coping was a protective factor for PTSD for adolescents in grade 9. Physical health condition and negative coping did not affect the risk of PTSD. Ego-resilience and social support also had no effect on PTSD.
Predictors for PTSD Using Multivariate Logistic Regression Among Three Grades
*P<0.05
† P<0.01
‡ P<0.001
Further analyses showed that all of the CMHP subgroups were associated with a higher prevalence of PTSD, with an OR of 1.90~24.58 (All P<0.05). The strongest associations between PTSD and paranoia, depression, and anxiety were identified.
DISCUSSION
Compared with several previous studies conducted within 2 years after the earthquake, the prevalence of PTSD (23.4%) in our study sample was not markedly lower. This massive and deadly earthquake might be prone to causing chronic or delayed PTSD in early adolescents. The long-term prevalence could also be due to that the participants in our study did not receive systematic psychological intervention (qualitative interview information). In a prior study of a catastrophic disaster, PTSD symptoms were worse at 3 years than at 1.5 years after the Spitak earthquake among untreated early adolescents.Reference Goenjian, Karayan and Pynoos 9 Chronic PTSD symptoms are likely to have future detrimental effects on these adolescents, including on the state of their conscience, their academic achievement, physical health, and the stability of their family and community life.
We found that, as the CMHPs of the surviving adolescents worsened, their risk of PTSD symptoms increased, with ORs of 3.98~17.67. Moreover, we found that all CMHP subgroups, such as anxiety, depression, and paranoia, were associated with a higher prevalence of PTSD. This finding indicates that the presence of CMHPs aggravated or impeded improvement in PTSD symptoms among adolescents.
Our results confirm the conclusions that a high self-esteem consistently buffers against the impact of negative stressors and that positive coping, a changeable protective factor, can lead to better mental health when adolescents are older in age.Reference Mann, Hosman, Schaalma and de Vries 5 , Reference Zhang, Liu and Jiang 10
We did not find an increased prevalence of PTSD among participants who experienced home damage or loss of possessions in the destruction. The main reason for this outcome is likely that the Chinese government initiated several positive measures to help earthquake-hit regions, such as providing central financial support and requiring 19 relatively wealthy provinces to provide one-for-one reconstruction in the destroyed areas. Post-disaster reconstruction was accomplished within 3 years. However, the surviving adolescents who experienced casualties in their family were still burdened by PTSD after even 4 years. More attention should be paid to these bereaved adolescents.
Our results suggest that younger adolescents were relatively safer from PTSD than older adolescents. The older participants in our study were in an earlier adolescent period (10-11 years) when the earthquake occurred in 2008. These findings imply that older adolescents, those who had started puberty, were more prone to a high rate of PTSD after natural disasters, potentially because their symptoms were compounded by other mood disorders.
LIMITATIONS
Some limitations applied to this study. First, the data were collected via self-reported questionnaires, thus overestimating or underestimating information could occur. Second, we did not have information about the PTSD rate of the surviving adolescents in grades 7 and 8; therefore, we could not determine the trajectory of PTSD among the same participants over time after the earthquake. Third, we did not evaluate other new traumatic events during our research period.
CONCLUSION
CMHPs hindered the improvement of PTSD symptoms. Adolescents who are older or who experienced casualties in their families need more psychological intervention for longer periods (at least 4 years) after an earthquake. Long-term interventions require further consideration of how to enhance self-esteem and promote positive coping strategies among surviving adolescents.
Acknowledgments
We thank all of the participants for taking part in this research. We also thank the teachers in the research schools for helping us address the survey.
Conflict of Interest Statement
The authors declare no conflict of interests.
Funding
This research is supported by the National Natural Science Foundation of China (NO: 30972546 and 81472994) and the fundamental Research Fund for Chinese Central Universities (NO: skqy201212).
