Posttraumatic stress disorder symptoms among child survivors of the 2008 Wenchuan earthquake: a comparison between Chinese ethnic Han and Hui groups

Procedures

This present research is a secondary analysis that draws upon a sub-sample from a large cross-sectional survey conducted in 2009 (one year after the Wenchuan earthquake). The survey included 21,652 children between 7 and 15 years old who were survivors of the Wenchuan earthquake and who lived in Qingchuan County, Guangyuan City, at the time of the survey. Qingchuan County was one of areas being most severely affected by the 2008 Wenchuan earthquake. The survey was commissioned by the local department of education to screen for mental health problems in order to provide the appropriate mental health services. The survey was administered in the school through face-to-face interviews by trained interviewers who were psychiatrists, psychologists, psychiatric nurses, or social workers. A detailed description of the methods and recruitment for the survey have been described in a previous work (Chen et al., 2017; Xu et al., 2012). Written informed consent was provided by the parents or legal guardians of the children during the data collection and the survey was performed in accordance with the Declaration of Helsinki. The ethics approval for the secondary analysis of the data for research purposes was obtained from the Medical Ethics Committee of West China Hospital (No. 2015-52).

Samples

In this secondary analysis study, we examined the data on children from two Hui Autonomous Townships of Qingchuan County. The sample selection process is illustrated in Fig. 1. The Hui Autonomous Townships are populated by Hui and Han people who have lived together (Zhou, 1996) under similar socioeconomic conditions. The two groups share similar macro-level environmental, spatial and socioeconomic conditions of the two ethnic groups. The children of the Hui and Han ethnic groups accounted for 99.9% (n = 2,287) of all of the children (n = 2,289) in the two autonomous townships. A total of 2,198 children (96.0%) completed the interview of survey. The Han people represent the largest ethnic group and the Hui are one of the main ethnic minorities in China. Due to their numbers in the affected regions, Hui people were the ethnic minority most heavily impacted by the Wenchuan earthquake in the autonomous townships. The survey included 426 Hui children aged between 7 to 15 who were living in Qingchuan County. A total of 247 lived in two Hui Autonomous Townships and the others (n = 179) resided in various parts of Qingchuan County. The total sample of this study consisted of 247 (11.2%) ethnic Hui children and 1951 (88.8%) ethnic Han children in the two autonomous townships. Data on the participants’ sex, age, place of residence, and resettlement following the earthquake are presented in Table 1.

Measurements

The demographic variables, including sex, age, place of residence (rural or urban), and resettlement following the earthquake, were collected by a self-administered questionnaire.

An earthquake-modified version of the PsySTART Rapid Triage System was used to assess the earthquake exposure experiences of children (Gurwitch et al., 2004; Pynoos, Schreiber & Steinberg, 2004; Thienkrua et al., 2006). This system included 11 yes/no questions adapted from an earthquake exposure scale informed by the Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) A-1 and A-2 criteria for PTSD. The system is composed of 6 items that serve to identify objective features of earthquake exposure (i.e. having experienced an earthquake, having been injured, having seen people die or being injured, having lost close family member or friend, having close family members or friends injured, having lost one’s home or important belongings), and 5 items that serve to identify subjective exposures (i.e. having felt unable to escape from the disaster, having felt panic of others, having felt extreme panic or fear, having been trapped for a longer time, having felt one’s own or a family member’s life to have been in danger) (Chen et al., 2017). Participants were instructed to answer “yes” or “no” to each of the items. All the children had experienced the Wenchuan earthquake, and all of the participants answered “yes” to the item of “having experienced an earthquake” in this study. Therefore, this item is not listed in Table 1 and was not considered in the subsequent analysis.

The adolescent version of the University of California, Los Angeles PTSD-Reaction Index (UCLA PTSD-RI) has been widely used to assess traumatized children following major disasters and catastrophic violence (Steinberg et al., 2004). The index has been translated into Chinese and was used to evaluate the symptoms of PTSD in our study. It contains 20 items that assess intrusion, avoidance/numbness, and arousal symptoms. Each item is rated on a 5-point scale ranging from “never” (0) to “most of the time” (4). The Cronbach’s α was 0.91, 0.79, 0.80 and 0.74 for the total scale, intrusion symptoms, avoidance/numbness symptoms and arousal symptoms, respectively (Chen, Hung & Lin, 2000; Chen et al., 2002; Zhang, 2006). If the frequency of any intrusion symptoms assessment item had a value of greater than or equal to 3, the intrusion symptom (B criteria) was positive (percentage of screening positive for B criteria). If the frequency of any three avoidance/numbness symptoms assessment items was greater than or equal to 3, the avoidance/numbness symptom (C criteria) was positive (percentage of screening positive for C criteria). If the frequency of any two arousal symptoms assessment item was greater than or equal to 3, the arousal symptom (D criteria) was positive (percentage of screening positive for D criteria). If the symptoms of a participant met all the B, C and D criteria, the subject met the diagnostic criteria of DSM-IV and was screened positive for PTSD in this study (Zhang, 2006).

The Junior Eysenck Personality Questionnaire (JEPQ) is a self-report questionnaire for children aged 7 to 15 (Claridge, 1977; Eysenck, Barrett & Saklofske, 2020). Because it was translated into Chinese and modified in the 1980s (Chen, 1983; Gong, 1984a), it has been widely used in China over the past 30 years (Gong, 1984b; Zhu et al., 1983). The Chinese version of the JEPQ, which was used to assess personality characteristics in our study, was comprised of 88 yes/no items (scored as “yes” = 1, “no” = 0) with four sub-scales: psychoticism/socialization (P), extraversion/introversion (E), neuroticism/stability (N) and lie (L). The Cronbach’s α coefficient of the P, E, N and L subscales and the JEPQ were 0.80, 0.83, 0.62 and 0.78, respectively (Gong, 1984b; Lyu, 2008). The score of each subscale was divided into five levels. For instance, the P subscale was divided into socialization (T < 38.5), moderate socialization (38.5 ≤ T < 43.3), intermediate (43.3 ≤ T ≤ 56.7), moderate psychoticism (56.7 < T ≤ 61.5) and psychoticism (T > 61.5).

Data analysis

A normal Q–Q plot was used to observe the normal distribution of the continuous variables. The t-test was used to compare the continuous variables with a normal distribution. The Mann–Whitney U test was used to compare the continuous variables that did not conform to a normal distribution. The chi-square test was used to examine the difference of categorical variables between the two ethnic groups. The PTSD symptoms which had statistical significance between the two ethnic groups were included as dependent variables in the subsequent regression analyses. The multiple linear regression was used to investigate the association between the ethnicity (independent variable) and the severity of PTSD symptoms based on the UCLA PTSD-RI score (dependent variable). Multiple logistic regression analysis was used to investigate the association between ethnicity (independent variable) and the percentage of screening positive for PTSD symptoms (dependent variable). The demographic variables, the JEPQ scores of the four personality traits, the objective earthquake exposure experiences and the subjective earthquake exposure experiences were included as independent variables step by step. The regression analysis yielded B or odds ratio with a 95% confidence interval (CI) for each independent variable. All tests were two-tailed and an alpha level of below 0.05 was considered to be statistically significant. All statistical analyses were performed using the SPSS 16.0.

Demographic

Sex and resettlement to another area after the earthquake were variables that differed significantly between the Hui and Han ethnic child groups (Table 1). Age differed significantly between the Hui (median: 12.0, interquartile range: 4.0) and Han (median: 12.0, interquartile range: 4.0) ethnic child groups (U = 216,585.50, p = 0.009), while age groups did not (Table 1).

Personality characteristics

The median (interquartile range, IQR) of the P (U = 234,192.00, p = 0.472), E (U = 232,213.50, p = 0.353), N (U = 237,763.00, p = 0.735), and L (U = 225,181.50, p = 0.093) subscales did not differ significantly between the ethnic Hui and Han child groups. Each subscale was divided into five levels and the distribution of the levels did not differ significantly between the ethnic Hui and Han child groups in each subscale of the JEPQ (Table 1).

Earthquake exposure experiences

Three subjective earthquake exposure experiences showed statistically significant differences between the Hui and Han ethnic child groups (Fig. 2): having been trapped for a longer time (Han: 755, 38.7%; Hui: 126, 51.0%), having felt unable to escape from the disaster (Han: 952, 48.8%; Hui: 151, 61.1%), and having felt extreme panic or fear (Han: 1587, 81.3%; Hui: 217, 87.9%). Two objective earthquake exposure experiences, including having lost a close family member or friend (Han: 283, 14.5%; Hui: 48, 19.4%) and having lost home or important belongings (Han: 1169, 59.9%; Hui: 170, 68.8%), yielded statistically significant differences between the ethnic Hui and Han child groups (Fig. 2A).

Han-Hui differences in PTSD symptoms

The average UCLA PTSD-RI total score for the ethnic Hui child group (27.01 ± 9.24) was significantly higher than that of the ethnic Han group (25.12 ± 9.17) (t = −3.05, p = 0.002), as were those for avoidance/numbness score (Hui: 10.02 ± 4.82; Han: 9.04 ± 4.60, t = −3.12, p = 0.002) and arousal scores (Hui: 9.36 ± 3.64; Han: 8.79 ± 3.42, t = −2.44, p = 0.015) (Fig. 2B). There was a significant difference in the percentage of screening positive for D criteria (arousal symptoms) between the ethnic Hui (48.6%, 95% CI [42.3–54.9%]) and Han (41.9%, 95% CI [39.7–44.1%]) child groups (χ² = 3.97, p = 0.046). The percentage of screening positive for B criteria (intrusion) (Hui: 45.7%, 95% CI [39.0–52.0%]; Han: 46.2%, 95% CI [44.0–48.0%], χ² = 0.02, p = 0.89), C criteria (avoidance/numbness) (Hui: 19.8%, 95% CI [15.0–25.0%]; Han: 15.5%, 95% CI [14.0–17.0%], χ² = 3.10, p = 0.08) and PTSD symptoms (Hui: 9.3%, 95% CI [5.7–13.0%]; Han: 8.0%, 95% CI [6.9–9.3%], χ² = 0.99, p = 0.32) did not differ significantly between the Hui and Han ethnic child groups (Fig. 2C).

Relationship between ethnic group and PTSD symptoms

We found a positive association between the ethnicity and the avoidance/numbness symptoms score (Tables 2 and 3). No association was observed between the ethnicity and the arousal symptoms score. The ethnic Hui minority was initially associated with a higher arousal symptoms score, higher PTSD total score, and higher percentage of screening positive for D criteria (arousal symptoms) following adjustment for demographic variables and personality traits (Models 2 and 3, Table 2). However, the significant association between the ethnicity and the percentage of screening positive for D criteria (arousal symptoms) (Model 1: OR = 1.31, 95% CI [1.00–1.71], p = 0.047) disappeared following further adjustment for objective earthquake exposure experiences (Model 4: OR = 1.32, 95% CI [1.00–1.75], p = 0.052). The significant associations between the ethnicity and the arousal symptoms score (Model 1: B = 0.57, 95% CI [0.11–1.02], p = 0.015) and PTSD total scores (Model 1: B = 1.89, 95% CI [0.67–3.10], p = 0.002) disappeared following further adjustment for subjective earthquake exposure experiences (Model 5: arousal symptoms, B = 0.41, 95% CI [−0.01 to 0.83], p = 0.056; PTSD, B = 1.00, 95% CI [−0.07 to 2.07], p = 0.066). Detailed information for Models 1 to 5 for significant PTSD symptoms is listed in the Supplemental Materials.

The associations between demographic variables, personality traits, earthquake exposure experiences and the PTSD symptoms are listed in Table 3. Sex was not related to the avoidance/numbness symptoms score, but was related to the arousal symptoms score. Age and the E subscale score were negatively associated with the avoidance/numbness symptoms score. The P, E, N and L subscale scores were positively associated with the avoidance/numbness symptoms score. Three objective earthquake exposure experiences (including lost home or important belongings, close family member or friend was injured, lost close family member or friend) and one subjective earthquake exposure experience (felt unable to escape from the disaster) were associated with the avoidance/numbness symptoms score.